
Why your user research roadmap runs behind the product roadmap
When user research doesn’t change the roadmap, the usual diagnosis is that it wasn’t persuasive enough. The report ran long, the insight lacked a story, or stakeholders didn’t value the work. A sharper readout cannot help a finding that arrives after the decision, because it was never competing on persuasion. I’d rather a team saw a rough finding in time than a polished one too late.
A user research roadmap tends to run behind the product roadmap, so a study reports after the money is committed and the sequence is set. The point where a decision stops being cheap to reverse is its commitment point, and findings fail around it in three ways. They arrive after it, reach people who cannot act on them, or are agreed and never shipped. A better report fixes none of the three.
Findings that arrive after the decision
Nielsen Norman Group’s 2026 guidance on getting research onto the roadmap says decisions are often half made by the time research arrives, and that a modest study that reports early can outdo a bigger one that reports after the plan is set. Teams usually test usability towards the end of a project, which limits the findings to incremental fixes and casts testing as refinement, according to a 2019 article on undervaluing user research. A finding that arrives after the commitment point can still change a label or a layout. It can’t change which problem the team chose to solve.
The gaps show up once a project is mapped. In workshops described in the 2019 Center Centre article, design leaders mapped a recent project on a timeline and marked every point where someone watched real users try the design. Once the gaps showed, every one of them concluded they weren’t doing enough research. My view is that more research helps only if it lands before the decision, because a team can test every month and still be late on every decision that matters.
Findings that reach people who cannot act on them
The people who hold the most sway over a project see its users least. The same 2019 Center Centre article groups the people on those project timelines into user experience professionals, other full-time staff, and occasional visitors such as executives, Legal and Compliance. The executives hold the budget and the sign-off, and the article describes them as insulated from the users.
A report cannot close that distance. Findings only help when stakeholders can absorb them, says research advice on report design, which rests on experience rather than a test of readers. What closes the distance is contact, and Nielsen Norman Group’s 2026 roadmap guidance recommends engaging stakeholders before pitching recommendations.
Leaders are also rewarded for shipping features and rarely penalised when they fail. That is why the gap stays open. A team rewarded for shipping has little reason to leave room for a finding that might slow it down.
Findings that are agreed and never shipped
In a Digital Equipment Corporation paper on the impact of inspections, specialists inspected 10 commercial software products in 1994 and 1995 and found 664 usability problems. First came the commitment. The development teams committed to fix 519 of the 664 problems, about four in five.
Next came the release. For six of the products, the authors could see the version that shipped. Five delivered what the team had committed to. The sixth had committed to 71% and delivered 37%.
So findings were lost in two places. About one problem in five was never committed to, and in one product the commitment was not kept. The reasons teams gave included too little time or resources, shipping being the top priority, and work pushed back to the next version. The study is from the mid-1990s and software has changed a lot since, but a gap between agreeing to a fix and shipping it is still worth checking for in any team.
Why the usual cost figures can’t support a business case
Late findings are usually defended with a number. Three figures circulate in the field, so we traced each one. Each may point to something real, but the evidence behind them is weaker than the way they’re usually quoted.
| Figure | Where it comes from | What it measures | Traceable |
|---|---|---|---|
| 100 times more to fix after release | A 2002 magazine article on usability returns citing a 1988 book and a 2001 International Business Machines (IBM) source. | The cost of fixing a defect late | No. A 2021 news report on the 100 times claim says its textbook cited IBM course notes, with no data behind the chart. |
| 135% average improvement | Nielsen Norman Group’s 2003 article on usability returns, from 42 website redesigns. A 2008 update on usability returns put it at 83%. | Change in usability metrics after a redesign | Yes, in both articles. |
| $1 returns $100 | A statistics roundup crediting a 2016 Forrester report. | Not stated | No. Forrester’s public page for the 2016 report shows a price of $1,495 and no such figure. |
The 135% figure needs a caveat. Nielsen Norman Group excluded five outliers from it, and its 2008 update calls the gap between the two surveys only marginally significant. A number that fell from 135% to 83% in about six years describes one period, not a constant.
None of the three measures what a late finding costs a team. A business case built on any of them rests on a number nobody can source.
You won’t get a figure like “100 times” from us, and you should be wary of anyone who offers one. What we can say is narrower. The more that has been built by the commitment point, the harder a finding is to act on, though nobody has measured by how much.
What a finding in time is worth
For our moderated usability testing and experimentation for Fortnum & Mason, testing showed customers struggling to see hamper contents on mobile listing pages, and a test that showed the number of items in each hamper lifted mobile conversion by 7.58%. A separate test of a single-column layout, aimed at bigger images, cut conversion by 15%. Because it was tested before rollout, the layout never reached every customer, and shipping it untested could have lost the business millions in revenue every year. To me, a loss avoided like that is worth as much as an uplift, even though it’s harder to show on a dashboard.
Our task-based usability testing for a luggage brand found that many participants missed the size selector in its dropdown. Those observations became ranked hypotheses for testing, as part of a programme of 19 tests. The research came before any design work, and the grid designs that followed were tested against the existing selector.
How to ask for a user research roadmap that can change the plan
Five requests put the research schedule in front of the decision. They are our own recommendations, drawn from the sources above.
- Check when the roadmap decision locks: get the planning calendar and the commitment point for each roadmap item, then push for fieldwork to finish before it.
- Ask what would change the plan: before fieldwork starts, get one sentence from the decision owner naming a result that would move the item. Add one naming a result that would not.
- Request a small study on time: a few sessions that report before the commitment point beat a large study that reports after it. Only the first can still change the decision.
- Require the evidence type: every finding should say if it came from observed behaviour, interviews, a survey or analytics, with counts and not percentages for small samples.
- Ask who owns each finding: name the person who decides on it and the date they decide, so the finding can’t sit in a deck.
Draw your last project as a timeline, then mark each date a decision owner or a researcher watched a real person use the design. Any stretch before the commitment point with no marks is where findings will arrive too late.
Which roadmap decisions can your research still change
Before you commission the next study, take the roadmap item it is meant to inform. How many weeks of delivery work would it cost to reverse that item, and on what date does that figure start to rise?
We plan each study back from the date a decision locks, and our user research service sets out how.
Frequently asked questions
What is user research?
User research is the study of how people behave and what they need when using a product or service. Methods include usability testing, interviews and surveys. The aim is to give a team evidence to decide with before it commits budget to a design or a feature.
Why conduct user research?
Teams conduct user research to find out what customers struggle with before they spend build time on a change. The reasoning is that a problem found in a handful of sessions costs less than the same problem found after release. The size of that difference varies by product, and no verified source we found gives a general multiple.
How to involve stakeholders in user research?
Invite the people who own the decision to watch live sessions, and agree before fieldwork what result would change their plan. Nielsen Norman Group’s 2026 guidance on getting research onto the roadmap recommends engaging stakeholders before pitching recommendations, though it rests on experience and not on a study. Our addition is the agreement on what would change the plan, because attendance alone doesn’t settle when a decision locks.
How much does user research cost per study?
No published source we found gives a verified typical price per study, because cost depends on the method, the number of participants and how they are recruited. The nearest measured figure comes from Nielsen Norman Group’s 2003 article on usability returns. It found usability work took 8% to 13% of project budgets across 863 design projects. That is a share of a project and not the price of one study.
How to present user research findings?
Lead with the decision the finding affects and state the type of evidence and the number of participants. Then give a recommendation with an owner and a date. The research vendor MeasuringU’s 2026 advice on report design adds layout tips such as a clear visual hierarchy and two or three colours. These rest on experience and not on a test of readers.
Sources
- Nielsen Norman Group, How to Get Research Recommendations on the Roadmap, Laura Klein, 29th May 2026.
- Nielsen Norman Group, Why Organizations Don’t Do User Research and How to Change That, Laura Klein, 22nd August 2025.
- Nielsen Norman Group, Strength of User Research Evidence, Jakob Nielsen, 14th April 2013.
- Nielsen Norman Group, Return on Investment for Usability, Jakob Nielsen, 6th January 2003.
- Nielsen Norman Group, Usability ROI Declining, But Still Strong, Jakob Nielsen, 21st January 2008.
- Nielsen Norman Group, Interpreting Contradictory UX Research Findings, Kate Moran, 27th January 2019.
- Center Centre, Undervaluing User Research is a Deadly Disease, Jared M. Spool, 19th November 2019.
- Jared M. Spool, How a Team Matures Its User Research Integration, 18th April 2019.
- MeasuringU (vendor research), Nine Design Fixes for Your UX Research Reports, Fernanda Villalobos and Jeff Sauro, 1st September 2026.
- Paul Sawyer, Alicia Flanders and Dennis Wixon, Digital Equipment Corporation, Making A Difference – The Impact of Inspections. The page states no venue or year.
- Aaron Marcus, The ROI of Usability, User Experience Magazine, Winter 2002, republished by UXPA 27th March 2013.
- The Register (secondary report), Everyone cites that bugs are 100x more expensive to fix in production research, 22nd July 2021.
- Hillel Wayne, Computer Things (secondary report), I ****ing hate Science, 19th July 2021.
- Forrester (analyst research, paywalled, figure not found), The Six Steps For Justifying Better UX, 28th December 2016.
- REO Digital case studies, IT Luggage and Fortnum & Mason.