ICON Public Limited Company (ICLR) Q4 2025 Earnings Report, Transcript and Summary

Good day, and thank you for standing by. Welcome to the ICON plc Q4 and Full Year 2025 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to hand the conference over to your speaker today, Kate Haven. Please go ahead.

Hello, and thank you for joining us today. I'm joined on the call by our CEO, Barry Balfe; and our CFO, Nigel Clerkin. I would like to note that this call is webcast and that there are slides available to download on our website to accompany today's call. Certain statements in today's call will be forward-looking statements. These statements are based on management's current expectations and information currently available, including current economic and industry conditions. Actual results may differ materially from those stated or implied by forward-looking statements due to risks and uncertainties associated with the company's business, and listeners are cautioned that forward-looking statements are not guarantees of future performance. Forward-looking statements are only as of the date they are made, and we do not undertake any obligation to update publicly any forward-looking statements, either as a result of new information, future events or otherwise. More information about the risks and uncertainties relating to these forward-looking statements may be found in SEC reports filed by the company, including the Form 20-F filed on May 27, 2026. This presentation includes selected non-GAAP financial measures, which Barry and Nigel will be referencing in their prepared remarks. For a presentation of the most directly comparable GAAP financial measures, please refer to the section of the press release dated May 27, 2026, titled Consolidated Statements of Operations. While non-GAAP financial measures are not superior to or a substitute for the comparable GAAP measures, we believe certain non-GAAP information is more useful to investors for historical comparison purposes. Included in the press release and the earnings slides, you will note a reconciliation of non-GAAP measures. Adjusted EBITDA, adjusted net income and adjusted diluted earnings per share exclude amortization, stock-based compensation, foreign currency gains and losses, restructuring, transaction integration-related and other adjustments, transaction-related financing costs, fair value movement on investments and equity, goodwill impairment, impairment of nonfinancial assets and their related taxation effect. In the interest of time, may ask participants to keep their questions to one each. I would now like to hand the call over to our CEO, Barry Balfe.

Thanks, Kate. Last night, we released our Q4 and full year 2025 financial results, our 2026 guidance and also reported the findings of the recent investigation into certain accounting practices and controls. We have a lot of ground to cover today, but before we begin, I want to take a moment to recognize the significant efforts of the teams across ICON in recent months, in particular, the dedicated team that supported the completion of the investigation, but also the 40,000 strong workforce that stayed focused on delivering best-in-class research, supporting sites and patients and delivering for customers. Throughout a challenging chapter for ICON, these teams exemplified our partnership mentality, and I'm grateful for their dedication and efforts towards advancing our mission. Now before turning to our results, I'd like to address the investigation directly. The process was initiated in October 2025 after the management team raised concerns to the Audit Committee of the Board. The Audit Committee initiated an investigation, which was conducted by external legal counsel and supported by forensic and technical accounting advisers. This was comprehensive in scope, assessing not only the revenue recognition practices in our full service businesses, but also areas including billing and recording of cash. The investigation determined that from quarter 3, 2023 to quarter 4, 2024, improper adjustments were made to the clinical services revenue of the company. This impacted the timing of revenue recognition, though not quantum. The company also identified errors in certain inputs related to revenue recognition, specifically estimated cost to complete, the assessment of realizable value and certain manual adjustments in respect of clinical trial services contracts covering the same period and into 2025. We also identified presentation issues with unbilled and unearned revenue for contract assets and liabilities eligible for offset were not fully identified. The issues identified resulted in an overstatement of $65 million or 0.8% of full year 2023 revenue and $93 million or 1.1% in full year 2024. There was no impact on our customers nor was there any impact on our reported cash flow. As part of the investigation, we identified material weaknesses in ICON's internal controls over financial reporting. Entity-level controls, including the tone from management were not sufficient to enforce the monitoring and maintenance of a proper control environment. And the company did not design and operate effective internal controls to prevent material errors in revenue and related accounts. Extensive measures have been taken to ensure the accuracy of our financial statements, and we are implementing a comprehensive remediation plan, which Nigel will discuss in detail. Myself and the rest of the management team take very seriously our obligation to maintain reliable, rigorous controls. We are reassured to have identified and addressed these issues swiftly and effectively, and we are committed to ensuring they do not recur. I'd now like to turn to our results. Having previously called out improved execution on our commercial strategy as a core priority, I'm very pleased with our strong commercial performance in quarter 4. Low double-digit increase in RFP flow, win rates up right across our business, gross bookings of $3.2 billion and significantly reduced cancellations combined to yield net bookings of $2.9 billion, an increase of 19% year-over-year. Importantly, our direct fee book-to-bill was in line with our overall reported book-to-bill of 1.36x, an improvement on the mix in recent quarters. Commercial excellence has been a key strategic focus across the organization, and we are seeing clear evidence of progress across a range of measures. While win rate improvement was broad-based across the business, I am particularly pleased with a 5-point sequential uptick in biotech win rates, a personal priority that I laid out in prior calls. More broadly, we saw solid traction across customer groups with no single award value above $150 million. A critical enabler of our success has been our ability to flexibly meet our customers' needs across both service, functional and hybrid models of development, particularly as their preferred models change over time. In quarter 4, we saw a solid contribution of awards from existing long-term partners alongside an increasing ramp from more recent large and midsized partnerships. Cancellations in the quarter were $365 million, down meaningfully from the elevated levels seen in quarter 2 and quarter 3 last year and were broadly balanced across customer groups. It's important to acknowledge that while we have made changes to how we capture cancellations, the improved quarter-over-quarter performance is evident under both new and old methodologies. As I committed previously, the change to cancellation and backlog methodologies provides for increased transparency by providing investors with enhanced visibility into intra-quarter dynamics that are relevant to assessing our current and future financial performance. Nigel will take you through the detail of the changes to our policies and resulting impact when he covers the financials in detail. In terms of financial results for quarter 4, we saw stronger-than-anticipated revenue, driven by a marked increase in pass-through revenue. This was partially offset by findings of the investigation. Specifically, the changes made to cost to complete and realizable value estimate in our full-service business impacted earnings by over [ $50 million ] in the quarter. After a thorough review process, we believe these changes appropriately reflect the expectations for future performance across full-service contracts. These dynamics significantly impacted margin performance in the quarter, resulting in an adjusted EBITDA margin of 15.5% in quarter 4. Moving to our outlook for 2026. We issued our full year financial guidance of revenue in the range of $7.85 billion to $8.15 billion and adjusted earnings per share in the range of $10 to $11. These ranges reflect the importance of appropriately conservative estimates at this time and sustained quarter-over-quarter and year-over-year improvements, especially with respect to earnings. The guidance range also reflects the divestiture of the Symphony Health business, which impacts full year revenue by approximately 2%. We expect a headwind to revenue this year due to the challenging bookings environment we experienced from 2024 through the first 3 quarters of 2025 and in particular, the elevated cancellation activity in recent quarters. Pass-through revenue is projected to continue at similar levels on a full year basis to 2025. As we indicated on our last earnings call, our 2026 margin profile will be impacted by business mix, specifically FSO/FSP dynamics and sustained levels of pass-through revenue. There is also an impact from pricing pressures from prior quarters as previously awarded projects convert from bookings into revenue. We will continue to work to offset these factors through efficiency gains from automation activity and advanced technology deployment in addition to overall cost management with a focus on optimizing resource cost and location based on customer requirements. In parallel, we will continue to invest in expertise, prioritizing high-growth businesses like labs and early phase as well as therapeutic areas with potential for accelerated growth, such as advanced hematological diseases and women's health. Our Functional Service business continues to perform strongly as we support a number of partners that have adopted hybrid development models. We anticipate that phased evolution of sourcing models will continue, positioning us well to benefit from this key element of our differentiated offering. While 2026 will be a year of navigating near-term headwinds, the leading indicators that we monitor, bookings momentum, pipeline quality and the maturation of key partnerships give us confidence in accelerating growth as we move towards 2027. Looking now at the broader environment, we have been encouraged by indications of strengthening demand over several quarters. Biotech funding has been positive with particularly strong capital generation in the last 2 quarters. There has been notable activity [Technical Difficulty] stage clinical programs, a key focus area for ICON Biotech. In large pharma, development spending has been supported by customers continuing to invest in their late-stage pipelines. Opportunity flow through quarter 4 last year increased in the low double digits on a trailing 12-month basis, led by activity in large pharma and particular strength in TA such as cardiometabolic and oncology. Across the business, we have seen the quality of opportunities improve through quarter 4 and indeed into this year with an increase in the average value of opportunities advancing to decision. These quarter 4 trends have sustained into 2026, and we expect that quarter 1 awards and cancels will be broadly in line with quarter 4. In addition, the commercial environment in quarter 2 is similarly encouraging. We look forward to reporting these Q1 and Q2 numbers in June and July, respectively. Moving on to strategy. Since [Technical Difficulty] I've been actively reviewing our portfolio to identify areas where we can generate the most value for our stakeholders. We are focused on opportunities where growth can be accelerated in priority areas of the business through investments in our people, capabilities and technology that will better enable our teams and further differentiate our offering. As a result, we have reallocated investments to our laboratory services business, increasing automation across our labs as well as expanding our testing menu, where we have recently added over 100 new biomarker assays. We also invested in the expansion of our early phase clinical footprint, opening a new purpose-built Phase I clinic in San Antonio, Texas with over 130 beds, along with satellite outpatient centers in Houston, Texas and Lawrence Campus. These facilities are specifically designed to support first-in-human studies as well as healthy participants and patient cohort trials and expand our capability in this high-growth area. We announced a partnership with Advarra, integrating ICON's technology with Advarra's systems across a broad network of research sites. By connecting workflows and data more effectively at the site level, we can support faster and more predictable trial execution, improve operational visibility for sponsors and reduce the burden for our site partners. This partnership will better enable sites to perform clinical research, accelerate study start-up and increase patient recruitment. Additionally, during quarter 2, we completed the divestment of Symphony Health to HealthVerity, a health care technology business with significant access to data assets across health care claims, EMR and pharmacy sources. ICON will retain access to an expanded pool of health care data assets without the need to own the assets outright, thus advancing our strategy in real-world data while reallocating capital to priority growth areas. We will also have an established partner, who is focused in this area and [Technical Difficulty] to navigate the inherent opportunities and potential risks that AI presents to the commercial data segment. Together, these moves reflect a sharpening of ICON's strategic focus, deliberate prioritization of high-growth opportunities and decisive management of the portfolio with disciplined allocation of capital. In parallel with reviewing the portfolio, we have been refining and progressing the company's AI strategy. Recent advances in the capabilities of large language models, in particular, have been rapid and are facilitating global businesses to move from AI's experimentation to AI as core infrastructure. While large-scale adoption across industries will be phased, the opportunities for drug development are relatively clear. In the first instance, the area with the single largest potential for transformation is in drug discovery. While it has not yet manifested in industry, we will see the emergence of tools that help to better design and synthesize new molecular effect on target diseases and disease pathways. The result will be a greater number of targets, increased predictability, lower failure rates in the clinic and reduced uncertainty. In the aggregate, these trends are net positive for society, for drug development and for CROs. For ICON, our focus is in 3 primary areas. In the first instance, ICON is building the intelligence layer that connects expertise, data and AI across the trial life cycle. This enables teams to turn information into knowledge and data into insights, allowing for better decisions faster. Examples include an integrated control tower for project teams that facilitates next best action. Secondly, there are a range of productivity gains to be found through AI-enabled automation as Agentic capabilities accelerate and improve high-volume, highly repeatable processes across our business. These agents [Technical Difficulty] human expertise to be redirected to higher-value activities. Examples include enterprise adoption of the deployment of digital assistants that support routine site queries. And thirdly, we continue to develop domain-specific agents that are embedded within clinical trial workflows. Our proprietary Orbis capability functions as an agent of agents that facilitate seamless navigation across disparate data sources. Our proprietary contracting agent accelerates study start-up and our new CRA agent will increase the time and expertise available for site management and patient recruitment. So while there's obviously potential for AI to dilute certain revenue streams over time, for example, the automation of clinical study reports or the reduction of human effort in programming, the opportunities presented by AI are likely to offset these risks. It's also worth noting that CROs like ICON with the necessary scale to develop industry-leading platforms and the expertise to leverage [Technical Difficulty] proportionately from this shift. Finally, a word on capital allocation. In short, our approach to capital allocation is consistent with 2025, disciplined, guided by a defined framework and with a clear priority to return capital to shareholders through share repurchases while continuing to invest in our capabilities. Let me now hand over to Nigel to take you through our results in further detail.

Thanks, Barry. Let me start with an overview of the remediation actions we are taking in the light of the investigation findings. Our plan is focused on 4 main areas: one, organizational and personnel changes in key roles and enhancements to our compliance programs; two, revised policies and procedures related to revenue recognition; three, training; and four, enhanced internal controls over manual adjustments. These actions are underway [Technical Difficulty] implemented in 2026. A full description of the material weaknesses and the remedial actions we are taking is included in our 20-F filing. Turning to the changes we have made to our backlog and cancellation policies. As Barry set out, these changes have been made in response to specific dynamics within our business that were influencing our backlog metrics. Our policy and approach to gross awards reporting will be consistent with prior periods, which recognizes awards upon written confirmation from our customers that have a defined value within the quarter of notification for those awards are expected to start generating revenue within 12 months. Additionally, each award must have evidence of sufficient funding to support the intended development program to be included in backlog. We believe this approach provides stakeholders the best visibility to our current business development performance regardless of timing related to contracting or [Technical Difficulty] with regard to cancellations, we have modified our policy such that reported quarterly cancellation amounts now reflect in-period contract cancellation notifications from customers in addition to studies that have been identified by management as at risk for cancellation. This change in policy will more accurately reflect current cancellation activity in comparison to our previous approach of reporting cancellations when only termination or study closeout agreements were finalized with customers for contracted studies, which can take a significant amount of time. Additionally, this change allows for adjustments for inactive or on hold studies that are unlikely to proceed, which did not occur under the previous policy. Our treatment of FSP awards and their recognition into backlog is unchanged from our previous policy, which includes the expected revenue under the award. [Technical Difficulty] The changes made to our policy resulted in an adjustment to our backlog of approximately $3.9 billion at October 1, 2025. We are confident that these changes provide investors enhanced transparency on awards, cancellations and our overall backlog, allowing for a more accurate assessment of our business. The unsatisfied performance obligation or our backlog in accordance with GAAP will continue to be reported on a quarterly basis, reflecting the total value of contracted awards adjusted for realizable value calculations in accordance with GAAP. Now let me turn to the financial results for the fourth quarter and full year 2025. Revenue in quarter 4 was $2.1 billion, representing a year-on-year increase of 2.5% and an increase of 1.3% on quarter 3 2025. For [Technical Difficulty] $5 billion, an increase of 0.8% over 2024. Our Q4 revenue was approximately $100 million higher than the expectations underpinning the midpoint of the Q4 guidance we provided at our Q3 results call last October. This mainly reflects 2 things. Firstly, pass-through revenues came in over $150 million higher than we had anticipated as the increasing proportion of pass-throughs we had seen through the year accelerated further in the fourth quarter. Second, in connection with the investigation, we performed a comprehensive review of our cost to complete estimates across the full service portfolio. This resulted in the correction of some errors in previous periods, but also led to updates to estimates that impacted our Q4 results and resulted in direct fee revenues for the quarter [Technical Difficulty] lower than previously anticipated. This mix shift had a consequent and significant impact on our Q4 EBITDA margin and adjusted earnings per share with both coming in materially lower than the Q4 guidance midpoint. Adjusted gross margin for the quarter was 23.7% and 27.1% for the year compared to 30.9% and 29.3% in quarter 4 2024 and full year 2024, respectively. Adjusted SG&A expense was $174.5 million in quarter 4 or 8.3% of revenue. For the full year, adjusted SG&A expense was $701.9 million or 8.5% of revenue. Adjusted EBITDA was $327.1 million [Technical Difficulty] -- this compares to $387.7 million in Q3 2025 and $455.9 million in Q4 2024. For the full year 2025, adjusted EBITDA totaled $1,530.7 million or 18.6% of revenue. This compares to $1,670.4 million or 20.4% of revenue in 2024. Adjusted net interest expense was $46.5 million for quarter 4 and $184.4 million for full year 2025. On a full year basis, net interest expense declined $20.7 million or 10.1%. The effective tax rate was 19% for the quarter. The full year 2025 adjusted [Technical Difficulty] with a similar rate expected for 2026. Adjusted net income for the quarter was $195.1 million, equating to adjusted earnings per share of $2.52. This compares with adjusted earnings per share of $3.86 in quarter 4 2024 and $3.20 in quarter 3 2025. For the year, we have recorded adjusted earnings per share of $12.53. This compares with $13.37 for full year 2024. U.S. GAAP income from operations amounted to $207.8 million or 9.8% of quarter 4 revenue. U.S. GAAP net income in quarter 4 was $149.2 million or $1.93 per diluted share per [Technical Difficulty] diluted share for the equivalent prior year period. For the year, we recorded U.S. GAAP net income per diluted share of $2.90. This compares with $8.90 in 2024. From a cash perspective, quarter 4 had cash from operating activities of $234.2 million. Capital expenditure was $59.3 million, resulting in free cash flow in the quarter of $174.8 million, bringing our total year-to-date free cash flow to $862 million. At December 31, 2025, cash totaled $647.3 million and debt totaled $3.4 billion, leaving a net debt position of $2.8 billion. This was broadly in line with net debt at September 30, [Technical Difficulty]. We ended the quarter with a leverage ratio of 1.8x net debt to adjusted trailing 12-month EBITDA. Our balance sheet position remains very strong, which affords us the flexibility to continue to strategically deploy capital. We are focused on an approach that balances further investment in our business as well as future growth, while also prioritizing our return of capital to shareholders. We made significant share repurchases in the year totaling $750 million at an average price of $167 per share. And with that, I believe we're ready to open it up for questions.

[Operator Instructions] We are now going to proceed with our first question. And our first question comes from the line of Elizabeth Anderson from Evercore ISI.

It's nice to be speaking with you guys again. Thanks so much for all the updates on the company. I guess one of the questions I have sort of a combo short, long-term question is that like short term, it seems like sort of the bookings looks great even on the alternate of the new standard. Can you talk about how much of that you think is sort of market-driven in terms of like the funding cycle starting to recover and maybe some idiosyncratic things going on customer versus any sort of new strategic initiatives on your part or any additional focus areas or share gains? And then two, I think one thing that broadly speaking, investors are wrestling with in terms of the CRO industry more broadly is sort of like what -- given all these kinds of push-pulls, how to think about the long-term growth rate. So I don't know, it's -- obviously, there's a lot of things that are in flux. I don't know if you're fully prepared to comment on that. But like how -- I guess, maybe even conceptually, how you would think about any of those longer-term pushes and pulls as well?

Thanks, Elizabeth. It's Barry here. I'll take the first one, and Nigel might want to expand on the second one. I guess the short answer is it's both. It's evident that there's an improvement in the underlying market conditions. You can see uptick in RFP flow. And it's not just quantitatively, it's qualitatively. The number of those proposals that are going to decision is higher. The proportion of those proposals that are ballparking or pricing exercises is lower. The win rates on those proposals are higher in the quarter. And we've been pretty consistent over the last 15 months or so about making a priority out of improved commercial execution. So good that there's superior flow, but I think the teams have done particularly well to execute on those. And as you know, we made a virtue out of saying we had great win rates in pharma, which I'm happy to say upticked in the quarter, but we needed to diversify our sales channels into those by cross-selling, things like labs and early phase services. So good to see that progress. But particularly in biotech, we wanted to see more of the market, which we had been doing, but we also wanted to see a win rate uptick, certainly up above the 30% range. And I was pretty open about that not having happened in quarter 2 and quarter 3. I'm pleased to say it did in quarter 4. I've indicated in the remarks today that we'll be giving you similar news around quarter 1, and we feel pretty good about quarter 2 based on where we are roughly mid-quarter. So I think it's a combination of both of the 2. On longer-term growth rates, sectorally, I think there's been a lot of noise in the commentary. And you'll forgive me if I don't give you '27 guidance 5 seconds after giving you '26 guidance. But I think the fundamentals are ultimately what will drive it. Do we see a sustained investment thesis from large pharma as they tackle an LOE cliff? Yes, we do. Do we see a normalization of biotech raise and deployment of capital? Yes, we do. And do we see broad improvements in the underlying dynamics that are germane to things like cancellations as well? Yes, we do. I'm sure we'll talk about AI at another point on the call. But I think as long as those underlying fundamentals remain strong and we see normalization in some of those disruptive forces, we should be set for more normalized growth rates as we move on.

[Operator Instructions] And this question comes from the line of Patrick Donnelly from Citi.

Barry, maybe one for you, just on the bookings updates here. Can you just talk about how stringent the bookings policy is going to be? I mean, obviously, you guys took a deep look at it here. You removed the $3.9 billion. Cancels were quite low. It's just hard to tell with -- given that $4 billion, the cancels, what's going where. But can you just talk about how you approach this bookings policy? Is the effort here to make that book-to-bill correlation to growth a little bit higher? I know that's been a pushback in the industry for a while. It seems like you guys want to have kind of the "cleanest" bookings policy here. So I would love just a little more color as to what went into this analysis. And then again, how to think about the cancels number given the backlog adjustment, what went where? It would be helpful just to get a little more color there.

Let me take the second one first, Patrick, because it's the easier of the 2. We gave you the cancellations under the new and the old methodology. So the underlying rate of cancellations dropped in the quarter in either methodology. That's nothing to do with the methodological adjustment. That's just what's going on in the market, which I think we'd heralded from a quarter or 2 ago as something we expected towards the back end of the year. So we'll let that one stand on its merits. The broader issue though was around intent. I think I've been pretty consistent in saying that there was a bolus of studies in the backlog that weren't performing as expected and that I wouldn't be doing my job in the first quarter and seat if I didn't look at whether or not our methodology was contributing to that. And I thought it was. The reality is when you look at that backlog adjustment, less than 4% of that adjustment came from awards that were made in 2025. In fact, more than 75% of it was 2023 or older. So the company had a long-standing policy of only taking cancels out of backlog when there was a written notification of cancellation or termination. And there's all sorts of things that can fall through the cracks in that context. I don't think that's the best way of doing things. We already gave a GAAP backlog that speaks to when things are contracted and how they're going to realize over time. I felt it was also important to give non-GAAP measures that gave relevant information about intra-quarter dynamics. So you can see what we were awarded during the quarter. You can see what was canceled during the quarter, not just papered as canceled. Either we became aware that a study was not going ahead because the customer notified us or frankly, we were no longer satisfied ourselves that it was going to go ahead. There are instances in the biotech community, for example, where if a study encounters long-term potentially existential delays, it's not often in the customers' interest to paper that officially for whatever commercial reasons they may have themselves. We don't want to wait for that. We want to make sure that investors have a solid understanding of the real intra-quarter dynamics and the quarter-over-quarter performance of the business. So that's really what was driving the change. And sorry, to close out on the answer, the answer is absolutely in terms of rigor. It will be applied to the letter of the law. As was the old policy, I just think the new policy is a better law to apply in the first place.

[Operator Instructions] And this question comes from the line of Ann Hynes from Mizuho Securities.

And can you let us know what EBITDA margin is implied in 2026 guidance? And should we assume from a modeling perspective that margins improve through the year. So maybe Q4 2026 margin is much higher than what was reported in Q4 '25?

Ann, it's Nigel. I'll take that one. And so Q4 '25, let's start there. Obviously, we walked you through where we ended up with an EBITDA margin of 15.5% for where we finished last year. When you look at the guidance for 2026, obviously, it's a range. So it will depend exactly where we come out in the range on both top line and bottom line. But just for simplicity, if you take the midpoint of that guidance, which would imply revenue of around $8 billion, and EBITDA of somewhere around $1.3 billion, that's an EBITDA margin for the year of approximately 16.5%. So roughly 1% higher for the year than where we were in Q4. Where we would see that building as we go through the year, broadly on the top line at this stage, we'd expect revenue to be broadly stable as we go sequentially through the year. But we would expect to see incremental improvement on that margin evolution and to end the year at an exit run rate then obviously that would be higher than that average 16.5% for the year. And likewise, similar dynamics in the EPS metric as well, where obviously Q4, we did $2.52, Q4 '25 that is. So we'd anticipate Q1 '26 broadly similar to that, and we should see growth as we go through the year and exiting obviously at a higher run rate heading into next year would be the anticipation.

[Operator Instructions] And this question comes from the line of Sean Dodge from BMO Capital Markets.

Maybe just on the revenue restatements. With the impact of 2025 being positive, and it looks like as of the third quarter of '25 and into Q4 kind of increasingly positive. Is there anything you can share just for like purposes of 2026, like what the impact of the restatement is going to be to this year? How big is the contribution lift to 2026 revenue you're getting from that kind of all else equal?

Sean, it's Nigel again. Let me take that. I would say, directionally, not huge. So I wouldn't look to that as a broad factor. When you look at the guidance we have given, again, at the midpoint, just to take the midpoint within the range, -- that suggests a revenue decrease of about 3%. Within that, there is actually an FX tailwind of about 1%. So on a constant currency basis, it's actually about a 4% decline. Roughly half of that is from the divestment of the Symphony Health business and the other half is underlying organic decline in revenue. As Barry noted as well, within that, our pass-through revenues are expected to be approximately flat year-over-year. So we are looking at an underlying decrease in our direct fee revenue through the course of the year.

[Operator Instructions] And this question comes from the line of Justin Bowers from DB.

And pardon the potential redundancy, some of the prepared remarks broke up. But Nigel, could you discuss the $50 million callout from 4Q and the impact on the direct fee revenue? And is that all -- is that sort of all catch-up cost? And then just wanted to clarify one of the comments in Q&A on the margins. Were you saying that 1Q jump-off point is going to be similar to 4Q '25 and then it will progress throughout the year?

Justin, yes, so me again. So on the first one, so again, as I mentioned, look, related to the investigation, we did conduct a broad review of our cost to complete estimates across all of the full service portfolio. So flowing out of that, we did update our estimates. In some cases, that did identify some errors that we corrected in the previous periods, and that's part of the restatement analysis that we've laid out. But in some cases, it did lead to changes in estimates in the fourth quarter for the future cost to complete across a number of studies. So that's what we're getting at there, Justin, it's around what is the expected cost to complete across that portfolio of studies from here through completion of the studies. And then on the second question on the margin evolution. So again, yes, we'd expect EBITDA margin to be broadly probably similar in Q1 to where it was in Q4 and then to progress through the course of the year and exit the year at a higher run rate than the average for the year. I don't know, if you wanted to add anything to that, Barry.

Yes. I would just point you, Justin, to the remarks I made when we launched the investigation and when we updated you guys on it -- for me, while the investigation focused on GAAP accounting, it was very, very important to look more broadly and make sure that all of the assumptions pertinent to revenue were on point, were solid and provided the correct baseline for go-forward revenue recognition. And there is a certain degree of subjectivity under ASC 606 in those long-term full-service clinical trial service agreements. So we did a very broad review either to identify errors or, frankly, assumptions that we felt just warranted being revisited. So that's really what the impact was on Q4. And we thought that was important to do this now, to do it once and to do it transparently.

[Operator Instructions] And this question comes from the line of Luke Sergott from Barclays.

I think like -- and this is not just with you guys, but I think a big question across CROs in general is just on the pass-through dynamic, I understand with the dynamics that you guys had in the [ idio ] stuff from 4Q and how that's going to shape out through the year. But I think right now, where investors are just kind of staying on the sidelines is like we don't have a lot of visibility into what the actual P&L looks like and what's coming in through the backlog. So anything you can give us from a pass-through perspective on how the bookings and the conversion is going to perform over the next 12 to 18 months, just gives us a little bit of visibility or comfort that like, all right, this jump-off point of 15.5% is where we can go and model from there. Otherwise, it's just going to be kind of finger in the wind, if you will, trying to figure this out.

So Luke, you have my sympathy. It's a confounded data set, and it can be difficult to extrapolate one line item like the impact of pass-throughs. You know what, when we gave guidance on earnings, much like when we gave an upper limit for the impact of the investigation, we choose our numbers carefully. We want to make sure that we give you guys a point from which you can bottom out and move on from there. And that's consistent and will continue to be the case across the board. But on pass-through significantly, I am mindful of the challenge you're describing. Pass-throughs aren't a factor in certain parts of the business. They're a large factor in other parts of the business, and they are volatile across those parts of the business. That is the nature of research. Pass-throughs are dominated in the full-service space by investigator grants in the same way that they're dominated in the lab space by things like lab kits and reagents and so on. And where and when and how patients are recruited, where and when and how they do visits, where and when and how they incur things like lab costs, is subject to the recruitment dynamics on these trials. And if that was perfectly predictable, honestly, we'd have put all the other CROs out of business already. There is a certain amount of volatility in there, and that remains. But to your point, that's why we called out the elevated level of pass-throughs, which is, to a certain degree, a function of the therapeutic mix in recent quarters. It's why we called out the fact that the 1.36 book-to-bill is broadly consistent, in fact, almost identical across a 605 and 606 basis in quarter 4. That's encouraging. That's certainly an improvement on recent mix. And we are open to providing more color on those direct fee and pass-through dynamics as we move forward where we think that's appropriate. So I can't give you a perfect model on that. There isn't one to give, but I do think that the numbers we've given you in terms of guidance we've given you from a position of confidence, and we are obviously mindful of things like the impact of pass-throughs on those numbers. Nigel?

Luke, I might just add as well that we fully, fully appreciate the challenge that, that presents, as Barry said. So look, we'll obviously try to provide as much commentary as we can around that to help you. Look, we've obviously laid out guidance for this year, and I've walked you through where we think the cadence of that is as we flow through the year. We've touched on before and Barry talked about the factors driving that EBITDA margin evolution this year versus last year. Equally, I do think it's important, Barry touched on, we do continue to focus, obviously, on being disciplined on cost control on where we're making targeted investments and so on. And also, given the commercial performance over the last few quarters, while it's obviously far too early to talk about guidance for 2027, we just got guidance for 2026. We are encouraged by what we've been seeing in terms of that commercial performance. And frankly, again, we've talked before about the operating leverage that exists in a business like this. And we are seeing, to some degree, our margin evolution this year being in part a function of negative operating leverage. The reverse should also be true as you get back to growth. So again, as we look towards the back end of this year and into next year, we'd hope to see that dynamic changing. Obviously, pass-throughs will be what they will be, but we're clearly much more focused on underlying profitability and driving progress in earnings per share and actual EBITDA dollars over time.

Our next question comes from the line of Eric Coldwell from Baird.

Unfortunately, you cut out on me a few times during the prepared commentary. I missed the dollar impact of the cost to complete estimate change in the fourth quarter and how you were guiding that impact to continue through 2026 and beyond. So if we could just get those numbers again would be great. And then how long would the cost to complete estimate change take to -- does it work through? Does it wind down over time as you work through this book of prior awards contracts underway where you've made the estimate change? Does it eventually go away? Or is this more of a structural or permanent adjustment that you expect would continue to be a rate limiter to margin in the distant future as compared to the past?

Both fair questions, Eric, and apologies if there was issues on the line. The impact to Q4 was north of $50 million. And the answer is those dollars get redistributed across the lifetime of those projects. These are not revenue dollars or margin dollars that disappear. They simply get rephased across what are relatively long-term contracts. So there was a follow-up question on what's the impact of the investigation on 2026. In the context of an $8 billion midpoint and you're talking about $50 million of impact being rephased over multiyear long-term contracts. The answer is the impact in any of those subsequent years is likely to be muted.

[Operator Instructions] And this question comes from the line of Jailendra Singh from Truist.

Barry, you called for it. So let's talk about AI. Clearly, this has been overhang for the group this year. I would love to get your thoughts on how you think about the AI impacting the industry in terms of pricing, in terms of in-sourcing versus outsourcing from pharma companies. Do you see AI as a net revenue tailwind or net revenue headwind for your business? And any color from your current conversation with pharma companies? Are they pushing for discounts on AI efficiencies or not? Just give us more color like how you think about the impact over the short term and also long term.

Happy to give you an update on that now, Jailendra, and thanks for the question. I think it's something we're going to return to in much more detail actually on subsequent calls. I think the industry has done a bad job speaking to both sides of the ledger. Until we acknowledge that we've managed to automate a huge amount of the human effort and things like CSR generation or that we're not going to be coding databases in 2030 the way we were in 2020, then I think it's difficult for everybody to hear the other side of the ledger. So of course, there will be areas where AI is net revenue dilutive to particular functions of the business. The 2 examples I gave you are not particularly material, but it's there. It's a real thing. That's not unique to drug development, though. I think we have to see that in the context of industrial scale adoption of large language models, in particular, that will help across a range of high-volume, high repeated processes in industry and across commerce generally. For drug development, you may have heard my earlier remarks, I think the biggest single tailwind to drug development will be when we take uncertainty out of the model, when we increase the number of shots on goal and the strike rate of those shots on goal in drug development, I think that reduced uncertainty increases capital flows into the space. It increases the number of targets. It increases the number of new medicines, and that's net good for just about everybody. Where we are today is if you think about the rate of improvement in the last, I mean, literally 9 to 12 months, you see a tremendous difference in the capability now. So we've talked openly about deploying now that which works now, experimenting now with that which may work tomorrow, but staying aligned with the really transformational changes that will require multiyear evolutions over time. So that's why we went out and we talked to the NVIDIA of this world about CPU to GPU migrations. That's why we work -- signed up a deal with Anthropic to embed Claude in certain core workflows and to enable our coders and developers to bring new agents to bear in our own environment. That's why simple things, simple productivity tools like Microsoft Teams with Cloud-enabled back ends are really, really important. But it comes back to what customers value. When we talked to customers about the CRA agent we developed, very honestly, we went out, we said, what's the best CRA agent in the market and should we buy it or license it. And the teams came back and said, with these new tools, we don't have to buy it because we can beat it. We can spin up a superior platform, have it active in 2 to 3 months and get customers on board, not just with co-development, but with adoption. What customers care about is, can you give me time back of those site-facing resources to spend on higher-value activities? I don't need them shuffling through paper or worrying about which element of SDV or SDR they need to do more of. I need them talking to investigators and coordinators about where the patients are, is the right study for that site? Are we reducing patient burden in such a way that we don't just do better at patient recruitment, we do it more predictably. The other element of your question, I'm a little bit amused at times about AI as the big driver of in-source, outsource dynamics. If you take a 30-year view, we always had project management, so did our customers and yet they outsourced. We always had monitoring, so did our customers and yet they outsourced. We always had clinical data science and so did our customers and yet they outsourced. We'll always have AI tools, so will our customers, and yet they will partner with companies like us to outsource and to in-source and to help them move forward. And this is the key point about these tools. They're based on what are largely becoming commercially available large language models, but the expertise required to turn data into insight, to turn information into knowledge is still based on that human capital. It's still based on drug development expertise and drug development operations expertise. So we can help customers to take time, to take cost, to take fixed cost and to take predictability or unpredictability out of the model. So I don't see it as a particular driver. I think we will see, and I said it in my prepared remarks, I think I might have said it in the PR, we will continue to see conversations back and forth about what the development of hybrid models look like. And very soon, I think we'll stop talking about hybrid models. We'll just call them models. I genuinely don't know any large pharma, who aren't doing some stuff in-house, some stuff in-sourced, some stuff outsourced. Our job is to optimize at the interface of those 3 models and to make sure that we create value by doing so.

[Operator Instructions] And this question comes from the line of David Windley from Jefferies.

Barry, I sure hope that last answer, thanks for that one, but that last answer sinks in. I sure hope it does. The question that I have for you is around, I'll call it, client traction. One of the concerns, I think, stemmed from the higher level of cancellations, other speculation about ICON's position at some large clients. Your book-to-bill in the fourth quarter and your call out on large pharma, in particular, makes it sound like those are either improving or perhaps never were as bad as we feared. And I wondered if you could elaborate on that a little bit. And relatedly, if I could sneak this in on the multiple questions on the $50 million those sound to me like essentially, you're changing your cost to complete estimate for the balance of the study in the context of assuming -- correct me if I'm wrong, assuming that, that change in the cost to complete is not going to be covered by change order value. So changes in the margin on those projects, but those changes in estimates would be overweighted to the period where you make those changes, so in the fourth quarter. Am I right there? Or is there some amount of this change in estimate being related to Olive branches that you're offering to clients related to certain projects. And again, getting back to my core question about the client relationships. Sorry for the convoluted 2 questions, but hopefully, you can answer those, please.

Dave, you were so nice about the AI question. I'm going to have to answer both of yours. Look, I actually hope your second question is the one that sinks in. No, it's nothing to do with all of branches. And yes, you understand it correctly. When you think about revenue recognition on a cost to complete model, the 2 things that you're measuring are what percentage of the current contract value you expect to realize and how much cost you expect to burn to drive that current scope. It doesn't take account of future out of scopes that may or may not be coming, that may or may not be expected in terms of upscopes. So yes, you're absolutely right. You get a disproportionate disruption in the immediate term and then you phase it out over the long term over the balance of the project. But in the context of a cleanup, in the context of surplus of caution to make sure that I'm not talking about revenue corrections forever more, I felt it was really, really important to throw the blanket wide and to go really, really deep. I cannot emphasize enough the extent to which we put operational teams and finance teams through the ringer to scrub every single one of those assumptions on every single full-service contract across the business. So yes, you are right. On client traction, I read these notes about, oh, is there a problem in large client S and large client B. And it's maybe the only time where I'm disappointed that we don't comment on single customer-specific dynamics because, frankly, they don't resonate for me. I think the market is broadly aware of ICON's penetration around large pharma. I think the market is broadly aware of what our incumbency looks like. And I think the numbers speak to progress capturing share and continuing to execute really, really well there. We know what the revenue drivers have been. We know what the margin drivers have been. Happy to talk about them more. But customer acquisition and customer retention has not been part of this story, and I don't anticipate that it will be part of this story. So I don't think our position is changing with respect to clients other than for the better I got to keep a few bits of powder dry for Q1 and Q2 Day, but we'll talk about new customer acquisition in the new year and some of the segments we've talked about, not just large and biotech, but also in the midsize, which is an important area for us. So I'm going to go with opt and see, not as bad as certain people seem to worry out loud, but that's not to be complacent, Dave. We've been really pleased with the level of engagement with our customers. I said in January of last year, it was probably February of last year on the call that a major priority for me was not just being the best delivery engine in the business, but being the best partner in the business. That's about being flexible in how you formulate and blend your capabilities to meet the needs of individual customers. That message continues to resonate. Where we need to get better now is making sure that we offer those customers access to the broadest array of ICON capabilities possible to make sure that we maximize the operational value to them and we maximize the degree of that value that we can capture in return.

[Operator Instructions] And this question comes from the line of Michael Cherny from Leerink Partners.

I just want to dive in, Nigel, if I can, on the margin side. As you think about the comments you made regarding the progression over the course of the year, how do you see the visibility maybe at the midpoint, maybe wherever in that margin progression? And how much of it is your expectations on mix shift versus your expectations on operational efficiencies, other components that I would say are more within your control?

Yes. Thanks, Mike. Good question. It's a bit of both, frankly. So obviously, Barry touched on the composition of the Q4 wins and being broadly similar, whether on a 605 or a 606 basis. That has been a shift, as you said, from the previous few quarters. So that's encouraging in terms of the quality of those wins. And again, we touched on, we will always continue to focus on operational efficiency. And so we do continue to expect to make further progress as we go through the year. So it is a bit of both, Mike. And again, look, we feel obviously good about the guidance range we put out and the likely cadence of that as we go through the year. I do think at this stage, revenues should be broadly stable quarter-over-quarter, but we should see margin progression as we go through the year and exiting on a higher run rate then going into next year.

I would second that, Mike, when we think about how we manage those things that are within our control, we structure our investments accordingly. We make sure that we're reallocating capital to the highest priorities, but also to ensure that this company isn't just doing what we said we'd do, we're doing it incrementally better on a quarter-over-quarter basis. And we very much had that in mind when we set the expectation of progression quarter-over-quarter and indeed into next year.

[Operator Instructions] And this question comes from the line of Max Smock from William Blair.

I wanted to follow-up on Dave's questions around competitive dynamics. One of the things that really stood out to me in your prepared remarks was your commentary around the win rate improvement that you saw, particularly among small biotech. So just hoping you can give us some more color on what you think is driving that improvement. Obviously, it sounds like you're benefiting from more of a concerted effort to see more opportunities in that space, but again, also winning significantly more of those opportunities. So just curious your thoughts on what is driving the latter there in particular.

Thanks, Max. I want to be careful on winning significantly more. I said from the middle of quarter 1, I wanted to win more. And we saw a 5-point jump in Q4. We didn't see progression on that number up to then, quite frankly. We were bidding on more, but we're winning a similar percentage of it. So progression in Q4 that was sustained in Q1 and it's looking pretty good for Q2. But I just want to be careful about talking about some sort of transformational change in the biotech win rate. It's an important step in the right direction. 5 points is quite a lot, but it's 5 points. To your broader question, I said when I came into the COO seat in January of 2025 that I thought ICON had an outstanding biotech offering, but it wasn't perhaps configured probably or properly. And I didn't think that the go-to-market story had been put together optimally and not just the story, but the package. So what have we done since? We've looked at the structure of our biotech organization to make sure that biotech customers aren't getting the flexibility they need. There's a trade-off in large CROs, who have pretentions of being leaders in the biotech space. If you overindustrialize process, you dehumanize the customer experience, and that's bad. And I've openly said, I think ICON maybe went a little bit too far in that direction in '23 and '24. So we've rightsized that. We brought back a lot of dedicated customer-dedicated focused headcount in areas like project management, in areas like medical affairs, in areas like reg, in areas like study start-up to make sure there's a highly personalized and expert-led service for our biotech customers. So some of that's organizational. Some of it, honestly, is personnel. We brought in some outstanding leaders from inside and outside ICON on the project delivery side and on the executive side to engage better, particularly in some key markets. You think about Boston, you think about the Bay, you think about China. We've done some really good work to put the right people in the right jobs. And that's also true in commercial. I'd have to give tremendous credit to our commercial teams. I think they've done really, really well to feed back to the organization what is required to win in my market and then to advocate for that change. In biotech, as in all markets, but particularly in biotech, the CRO game is largely predicated on being a better advocate for your customers in-house than it is for being a better advocate for your customer or for your company within your customers. So there are some of the big ones. The other ones, very honestly, is about choosing where to play. We said we wanted to bid on more of the market, but we also need to prioritize those areas of the market where we want to win. Within that increased win rate, the average size of opportunity that we're winning is going up. So we're winning a higher proportion of the dollars even than we are a proportion of the proposals. And that says you're getting better at winning the ones that matter. In fact, the number and proportion and value of what we call bolder opportunities, north of $50 million increased significantly in quarter 4. And that's important to me. It's all well and good to win 1 in 3x or 1 in 2x or 2/3 of the time in pharma, but you got to make sure you're winning where you want to win. So investing in the right therapeutic, operational, executive and commercial expertise is key to being able to make that a very strategic strike on the opportunities you need to win rather than treating them all as equal. So there's a lot going on there, Max, but there are some of the key themes.

[Operator Instructions] And this question comes from the line of Michael Ryskin from Bank of America.

Great. Sorry, the audio was cutting in and out. So I apologize if this is beating a dead horse, but I want to go back to the 4Q margin and EPS number and sort of make sure I'm thinking through that correctly. I appreciate the color, but that's a jumping off point for 2026. You talked about going from 15.5% in the fourth quarter to 16.5% for '26 and sort of like the progression through the year as you go through that. But that's such a big departure from the margins we have for prior years, even post restatement, I want to make sure I understand that correctly. So you called out the $50 million in cost to complete. You also have the higher pass-through revenue mix. So question one, I guess, is, were those the only 2 factors impacting 4Q margins? And then the second question, I guess, a bigger part of that is if this cost to complete adjustment that you're sort of implementing in terms of your estimate and your review process in terms of setting more appropriate expectations, like should the margin in prior periods have been lower as well? I guess I'm trying to get at is like what is the more appropriate underlying margin for ICON? Because if that is the case that you're being too aggressive in some of those assumptions, shouldn't prior period margins have been lower, too? Just to get to that point, like what is the more appropriate underlying.

Yes. Michael, it's Nigel. I'll take that. And look, yes, we did get a lot of reports going through the call that what there did seem to be some interference with the landline we were using. So we actually redialed in on Kate's cell phone here for the Q&A. So anyway, it was what it was through the call. But just to repeat on Q4 specifically, what we did say in the prepared remarks was Q4, there were 2 impacts on the revenue line that then also impacted on the margin and EPS. One was pass-through revenues did come in north of $150 million higher than we had anticipated. And then on the other hand, direct fees were impacted by north of $50 million from the cost to complete estimate changes that we talked about. So just for the numbers. In terms of your comment on the back periods, no, obviously, we've gone through a comprehensive review of the historic periods. We have corrected the errors that were there. And so the reported margins that you see for the previous periods are the reported margins for those periods. When you look into this year, what we're seeing, again, just to go over again, we -- it's the factors we spoke about before. Firstly, pass-throughs obviously ended the year higher and will be broadly similar in 2026 as 2025. So pass-through as a proportion of the overall revenue mix is a factor versus previous years, for example. And then we've talked about the mix within the direct fee decline. So again, I talked about overall on an organic basis, revenue being down about 2% year-over-year with pass-throughs flattish. That is obviously all in direct fee, and that's a factor. Then you have the mix effects between FSO versus FSP. That's obviously an impact as well. And then we have touched on the pricing pressures, et cetera. So that run rate that we see for this year, which, again, we've put out a range in guidance. So the midpoint, I just picked for ease of reference, is about 1% higher at the midpoint versus 2025. We do expect to end the year within that at a higher run rate than that average for the year. And again, getting back into the outlook for next year, it is obviously early to call '27. But the commercial traction we're seeing would encourage us that in terms of the growth dynamics as we head into the back end of this year and into next year. So we should get back into operating leverage and enhanced margins as we go through into '27 and beyond. But hopefully, that helps.

[Operator Instructions] And this question comes from the line of Casey Woodring from JPMorgan.

Can you guys just hit on the pricing in the quarter and how that trended both in RFPs and awards? And then can you just walk us through the decision to keep the treatment of gross awards projected but not contracted and backlog the same? And just like what your updated thoughts are in terms of that visibility into, I think it's close to $7 billion of difference between the $22 billion in reported backlog and the $15 billion of unsatisfied performance obligations.

I'll take the second one first, Casey, if that's okay. We're winning about $3 billion a quarter, and it takes a couple of quarters to get these contracts signed. So that delta is almost exactly where you'd expect it to be. I mean there's not huge mystery to that, to be honest with you. The reason for not changing the inclusion criteria is because I don't think there's anything wrong with them. In the context of already giving the GAAP numbers, if we only give contracted numbers, you're going to see everything in lag. You're going to see everything in disproportionate lag at different times of the year where there's seasonality in contract signatures based on customer budgets, et cetera. And it will become less transparent for investors. Investors will know less. When we did the adjustment, I actually went back and had a look and said, was there anything in there that should have failed at the point in which it was included in backlog? And the answer, honestly is no, not materially so. You can always argue there was 1 or 2 subjective ones 4 or 5 years ago, but we're really, really careful about what goes into backlog. The difference is that when life happens to an opportunity, I don't think the old methodology was sufficiently discerning in when to remove them. And that's what we've corrected. That will be applied to the letter of the law. And I firmly believe that is by far the most transparent of the methodologies of anybody in the space. You can simply tell more about what's going on with gross bookings, with cancels, with net bookings and with contracting cycles. So that's the answer on that one. On pricing, I don't think there's anything material to note about pricing in the quarter. I think as an industry, we saw some volatility in '24, maybe into early '25, and then that's been largely stable ever since. What we are talking about is a period of time where those awards at those new levels, and we're talking about pricing differences at the margin here, but it does take time for them to equalize into the P&L and find some equilibrium. So that's what I think we're seeing play out. But the factors that Nigel described are much more relevant to pricing -- or to margin rather in the quarter. It's about mix. It's about the investigation. It's about FSO, FSP dynamics. It's about pass-throughs. It's all of these other things. So nothing material to report on pricing other than to say there have been times in the last 2 years where large customers have come and said, in this part of the portfolio, we want price point X or else. And we've chosen else. There have been times when even in large functional partnerships, people will come and say, we think your pricing in country X can be undercut by 20%. And the answer is a very candid and civil, if you could do that, we think you should. We just don't think you can do that. And there have been times where they've executed on that option. And within those, there's been some times where they've come back to us and said, [indiscernible], that wasn't doable. And we've actually done okay on those partnerships, too. So nothing new in that, but I think price discipline is going to be central to what we do on a go-forward basis. We are in the business of creating value for customers, not in the business of racing to the bottom on a sticker price.

[Operator Instructions] And this question comes from the line of Lucas Romanski from TD Cowen.

This is Lucas on for Charles Rhyee. I wanted to ask about the $3.9 billion adjustment to the backlog. How much of this adjustment is related to the new criteria that canceled the project when management makes the determination that the study is at risk as well as how much comes from the new criteria that contracts that are deemed inactive? And then digging in on the criteria for studies that have been identified as being at risk. Can you provide more detail on how this methodology works in practice and how that determination will be made?

Yes. So the $3.9 billion, Lucas, if you look at the difference between the 2 methodologies, and I think we included a table in the slides that we provided to accompany the call that calls us that the delta is where we'd either been notified that a study wasn't going ahead, but it hadn't been officially papered or where we simply didn't believe it was going ahead. And that might be because it was active and became inactive or it may simply be awards that never started. So I don't have a breakout of that for you. It's pretty broad-based. I think the important point to note is less than 4% of it was from 2025 awards and more than 75% of it, certainly more than 74% of it was 2023 or earlier. I think that is important to note. Going back to the second part of your question, I sort of answered it in the first. If a customer comes to us and says, a study that we awarded in January that was due to start in July is now going to go on 12-month hold. Well, now it no longer meets a criteria for starting within the next 12 months, and I'm going to take it out under the new methodology. If a customer comes to me and says, hey, we had the funding, but we've decided to reallocate it to another program and now we're looking for funding for that program, then it no longer meets the award criteria to sit in backlog. If a customer comes to us and says we've had manufacturing issues [indiscernible] and we don't know if we can run the study, then it no longer meets the criteria. And as I say the last part, yes, there's some discretion here, but it's only discretion in the conservative direction. If a customer goes quiet on us, if a customer doesn't have definitive answers, but we no longer have evidence that the study meets the criteria it needs to meet to be in backlog, then we will unilaterally take it out, notwithstanding an official notification from a customer. So I hope I've addressed the nature of your question there, but by all means, please let us know...

We are now going to take our last question. And this question comes from the line of Ryan Halsted from RBC.

My question is just on the burn rate. Q4 burn rate was much higher than trend and kind of higher than maybe some of your peers. But can you just elaborate on this 10% burn rate? Is it sustainable? Or how much of that is derived from maybe the updated backlog policy changes and/or the pass-through dynamic? Any clarity on that would be helpful.

That's a good question, Ryan. A couple of points, maybe 3. The first one is, obviously, it's a function of different things. When you think about the backlog adjustment being a function of historic awards that were sitting in backlog at 0 or close to 0 burn rate, then it stands to reason that by removing them, we should return to more normal historical burn rates north of 9%, which is what we saw in the quarter. I think that's important. When you think about it going forward, yes, our burn rate might be a little bit higher than others, but things like FSP, we only take 12 months into backlog. I know others take 5-year MFAs at estimated revenues into backlog. We don't do that. We only take 12 months. So the largest FSP business in the industry is sitting in there burning at about 25% per quarter, right? So there's a significant uptick in there. And I think that's notable and important. The other thing about the go forward, yes, there's nothing inherently there that I think should slow it down. The only thing is when you have very strong book-to-bill, you are going to be adding to backlog faster than you're burning it. So if you're taking along at a 1.0 book-to-bill, that's one thing. If you're taking along at a 1.36 or even a 1.4 book-to-bill, then you're going to see, over time, some volatility in that backlog number. But organically, the main drivers here are the simple weight of math in terms of what went on with the backlog, how fast we're burning the studies, the nature of our business and book-to-bill on a go-forward basis will be what determines it for the most part.

There are no further questions for today. I will hand the call back to Barry Wolfe for closing remarks.

Well, I'll just close by thanking you all for staying with us. I do apologize if there were issues on the line. We'll make sure to get a transcript out to you guys so you can follow on the prepared remarks. And thank you for staying on what was something of a marathon call. I know there was a lot of ground to cover today. We wanted to give as much granularity and as much transparency as we can. And we do have the opportunity to come back with Q1 numbers in June and Q2 numbers in July. So more than happy to deep dive on some of those areas of interest that we probably covered in summary form today. So thank you. We appreciate it, and have a good day.

Thank you. This concludes today's conference call. Thank you for participating. You may now disconnect.