UX Research ROI: How to Convince Your Management

For a client in the telecoms sector, UX metrics eventually became a factor in the executive board’s bonuses. Not overnight, but after nearly a decade of properly implemented tracking. That is the moment when UX truly takes hold in a company: when a research figure influences senior management’s pay. No pitch deck in the world can achieve what this single integration has achieved.

That’s precisely why I’m annoyed by the figure that crops up in almost every argument for a bigger research budget: every euro invested in UX yields a hundred in return, usually with the addendum ‘according to Forrester’. If you follow it up, after three or four clicks you end up at an article that quotes another article, which in turn refers to a paid report from 2016 that hardly anyone has read. Nobody who seriously wants to demonstrate the value of UX research needs this figure. It’s actually harmful, because it collapses at the first sign of critical questioning and drags the whole argument down with it.

What does hold up, however, can be substantiated. And it starts with a better question: What aspects of the value of UX research are truly verifiable, and what is simply repeated over and over again?

Key takeaways

• The ‘1 euro in, 100 euros out’ figure for UX still cannot be traced back to any verifiable study.

• The stronger argument for management isn’t about ‘promising a return on investment’, but about ‘highlighting the risks’.

• On average, five test users identify around 85 per cent of usability issues – a finding scientifically proven since 1993.

• According to McKinsey, companies with strong design capabilities grow significantly faster than their competitors, but ‘design’ is not the same as ‘UX research’.

• Even the Nielsen Norman Group’s own collection of ROI case studies, upon closer inspection, only partially delivers on its promises.

• In-house data from support, product and sales is more convincing than any external industry figures.

• One telecoms client has embedded UX metrics for almost a decade in such a way that they were factored into executive board bonuses.

Where the 100x figure comes from, and why the trail goes cold

The most common version goes like this: Forrester found that every dollar invested in UX yields a hundred dollars in return – a return on investment of 9,900 per cent. This figure has been passed around in blog articles, LinkedIn posts and pitch decks for years. Almost always without a link to the original source.

The most likely origin is a 2016 Forrester report entitled ‘The Six Steps For Justifying Better UX’. The report costs just under $1,500, meaning that most people who cite the figure have never actually seen the original. A second possible source is a statement by Dr Susan Weinschenk, according to which it is a hundred times more expensive to fix a bug after development than before, which in turn loosely stems from older cost models in software engineering from the 1980s. Both leads end in the same dead end: a plausible basic idea, but no study that one can read and verify for oneself.

This is not an isolated case in the IT world. The Standish Group’s CHAOS Report on failed IT projects has been cited in almost every project management presentation since the 1990s. In 2006, Robert Glass demonstrated in detail in the Communications of the ACM that the report’s methodology was never disclosed and that the figures are significantly less reliable than their widespread use would suggest. If even such a well-established citation fails to stand up to academic scrutiny, it is worth taking another look at one’s own favourite figures as well.

What is actually reliable in UX research

Two things can actually be substantiated, albeit in a different way to what the classic pitch deck suggests.

Small sample sizes are more robust than many people think

A common objection to qualitative UX research is that five or eight users is far too few to prove anything. This is precisely what Jakob Nielsen, together with Thomas K. Landauer, investigated as early as 1993 in a mathematical model published at the ACM’s CHI conference. Their finding: with five test users, on average around 85 per cent of existing usability problems can be uncovered. Whilst more participants do yield further findings, the marginal benefit decreases sharply.

This figure has been one of the most frequently cited in usability research for over thirty years, and it has also been the subject of critical debate ever since. Some studies show lower hit rates for more complex systems. Nevertheless, the core message remains valid: qualitative research with small sample sizes is not a stopgap solution, but a methodologically sound approach, provided the quality of the execution is right. This is particularly important when a product owner asks why you don’t simply survey 500 people to make a decision.

Design pays off, just not as ‘UX Research’ on its own

In its 2018 study ‘The Business Value of Design’, McKinsey examined 300 listed companies over a five-year period across three sectors, recording over 100,000 design decisions. The result: companies in the top quartile of the resulting Design Value Index saw revenue growth 32 percentage points higher than their industry competitors and achieved a return on equity 56 percentage points higher.

This is one of the most robust figures available on the subject. However, there is a caveat: McKinsey measures ‘design’ in the broadest sense, including product design, service design and organisational structure. UX research is part of this, not the whole picture. Anyone who presents this figure at face value as ‘research generates 32 per cent more revenue’ is making precisely the mistake we criticised at the start of this article: taking a figure out of the context in which it was produced.

Even the industry’s own benchmark is only partially accurate

The Nielsen Norman Group, one of the most respected names in the field, has been compiling a collection of case studies on usability ROI for years. In 2019, Ross Unger examined this collection in detail, case by case, for the trade magazine Boxes and Arrows. His findings were mixed. Some cases are genuinely well documented, such as those involving Performance Bikes, Macys.com or eBags, where usability measures and financial outcomes can be clearly linked. In a large proportion of the remaining cases, however, something crucial is missing: the cost of the measure itself. A hundred per cent increase in turnover sounds impressive, but says nothing about the actual return on investment as long as no one offsets it against the amount invested.

This is not a criticism of NN/g. Rather, it shows how difficult it is to obtain sound evidence of ROI in practice, even for an organisation with decades of experience and open access to its own customer data. Anyone building an internal business case themselves should be aware of this benchmark before applying it to others.

Why UX research is not a revenue promise, but risk management

Here is the real point of this article. The strongest way to justify the value of research is not the promise of a return on investment. It is making a risk visible. A product that fails to meet the needs of its target audience is not simply a missed opportunity. It is an active cost factor, on three levels simultaneously.

Firstly, development costs: time and budget are poured into features that nobody uses. Secondly, operational costs: every avoidable usability issue generates support enquiries that cost money, enquiry by enquiry. Thirdly, and most costly of all, trust costs: customers who have had a bad experience with a product once rarely return a second time, and this lost trust is almost impossible to regain.

This framework has an advantage that no ROI figure can offer: it cannot be refuted as soon as someone in management does a quick calculation. A return-on-investment figure can be questioned, scaled down, or countered with another study. But no one can seriously claim that a product developed without any user research is developed just as reliably as one that is. The question is no longer whether research reduces risk, but simply by how much and in what areas. And this question can be answered using the company’s own data, rather than importing it from an external industry study.

This shifts the whole conversation. Away from “Prove to me that it’s worth it” and towards “How much risk are we prepared to take?”. This is a question that decision-makers in every other area of their business are familiar with, from IT security to product liability. UX research thus ties in with a way of thinking that has long been established in management. We have described in more detail why this shift in language is worthwhile in ‘UX Research as Risk Management’.

In my work as a UX consultant, I regularly find that this shift in perspective acts as a door-opener. A CTO who is rightly sceptical of the 9,900 per cent figure listens differently when they hear ‘risk avoided’, because they recognise the concept from their own project management experience.

Why the telecoms example remains the exception

The case from the introduction works because the company was willing and able to stick with a measurement system for years. Two measurements a year, methodologically sound, for almost a decade. As soon as a figure affects one’s own salary, it is taken seriously, and over time this seriousness develops into a mindset that no longer treats UX as an optional extra.

However, most companies never reach this point. Not because they lack the will, but because the system behind it is rarely set up properly.

What most companies lack

The difference almost always lies in the measurement points. Effective UX tracking requires small, targeted surveys at the exact ‘moment of truth’ in the customer journey – the point where customers can directly describe their experience. Tracking is needed within websites and software to reveal what people actually do, not just what they say in a survey. And a structured identifier is required to link this qualitative data with business data.

The final step is the most difficult. Setting up surveys is feasible. Setting up tracking is feasible. The real hurdle is linking the two cleanly and permanently, so that a robust system is created rather than a collection of isolated data silos. Anyone who overcomes this hurdle will have a business case that is virtually impossible to refute, because it is derived directly from their own figures.

How to build an internal business case based on your own data

A business case based on external industry figures will fall apart at the first sign of critical questioning. One based on the company’s own data will hold up for much longer. Here’s how to go about it:

1. Gather what’s already there. Support tickets, checkout abandonment rates, sales enquiries, complaints. This data already exists in most companies; it’s just rarely linked to UX issues.

2. Work out the costs of the current situation. How many support enquiries are caused by a specific interface problem, and what is the average cost per enquiry? The support team usually knows this figure, not the UX team.

3. Conduct a small study before calling for a large one. Five to eight users, a clearly defined problem, a tangible result. The Nielsen-Landauer figure mentioned above provides the methodological backing for this.

4. Speak the language of the relevant stakeholders. A CTO thinks in terms of development effort and technical debt; a sales manager in terms of conversion rates; a finance department in terms of costs avoided. The same study can be framed differently for all three.

5. Document the results before moving on to the next study. A single, well-documented internal case study carries more weight than ten cited industry studies, because nobody within your own organisation will question its origin.

Which UX metrics really resonate with management

Not every UX metric is equally persuasive at board level. The System Usability Scale and task success rate are useful for the UX team itself, but mean little to a CFO. More effective are metrics that can be directly translated into familiar business indicators: abandonment rate in the payment process, average processing time for a support enquiry, time to first successful use of a new feature.

The difference lies not in the quality of the metric, but in how it is interpreted. A task success rate of 62 per cent is abstract in itself. The same figure, interpreted as ‘38 per cent of users abandon the process and contact support instead’, is immediately understandable to every level of management because it relates to a familiar cost centre. We explore how UX metrics can be used strategically in the workshop ‘Efficient Use of UX Metrics and Strategic Data Utilisation’.

Conclusion

The ‘100x ROI’ figure for UX research will continue to feature in pitch decks for a long time to come, simply because it is straightforward and catchy. It is not robust. What does hold up, however, is a robust methodology that can be scientifically validated, studies with transparent methodologies such as the McKinsey Design Value Index – when properly contextualised – and, above all, the company’s own data.

The most important step, however, is not a change in figures but a shift in perspective. Those who stop selling research as a revenue promise and start positioning it as risk management will have a conversation that resonates with management. Not ‘Prove to me that it’s worth it’, but ‘How much risk are we prepared to take?’. Anyone fighting internally for a larger research budget is better off with this framework and their own small, clearly documented case study than with yet another external study.

In our workshop ‘UX as Risk Management’, we’ll show you how UX can be consistently positioned as a risk factor at the decision-making level.

Would you like a template to help you build your own internal business case? Sign up to our newsletter, where we’ll be sharing it in the coming weeks.

Frequently Asked Questions

Is it true that every euro spent on UX yields a return of one hundred euros?

This figure has been circulating for years, but cannot be traced back to a verifiable original study. The presumed source is a paid Forrester report from 2016, the methodology of which is not publicly available. More reliable are more specific studies such as the McKinsey Design Value Index, although it should be noted that these define ‘design’ more broadly than ‘UX research’.

How do I make the case for a UX budget to management if I cannot cite any major studies?

The most effective approach is to use your own company data: support costs, abandonment rates, rework effort. It also helps to frame the argument in terms of ‘avoiding risk’ rather than ‘promising a return’, as this cannot be undermined by a single questionable industry figure.

Is a study with five users really enough?

For most qualitative questions, yes. The mathematical model developed by Nielsen and Landauer in 1993 shows that five test users uncover, on average, around 85 per cent of existing usability problems. Larger sample sizes are required for quantitative statements with statistical significance, but that is a different matter.

Do the UX metrics that really matter depend on who you’re talking to?

Yes, and that’s precisely what’s often overlooked. The same metric will convince a finance department differently from a product team. What works is consistently translating a UX metric into an already familiar business metric, rather than presenting it in isolation.

What does a UX tracking system need to really work?

Three building blocks: surveys at key moments in the customer journey, behavioural tracking on websites and in software, and a structural identifier that links both to business data. In practice, the last point is the most difficult and, at the same time, the most important.