During my studies, the NHS was a hot topic for academics and transformation managers due to its high complexity and profile.

In March 2025, the National Audit Office published an assessment of three NHS transformation programmes worth over £3 billion [1]. One programme broadly worked but missed its targets, and two fell well short: surgical hubs delivered 48% less additional activity than planned, and a programme to cut follow-up appointments by 25% achieved 0.1%. The one that did work, or at least showed promise, earned clinical credibility before investment — its approach was grounded in an independent clinical review and built on securing clinical support. Credibility wasn’t a silver bullet — one of those two had clinical buy-in and still missed. But it was the one thing the programme that worked had, and the worst performer never did — its headline target had no supporting evidence and its own royal college declined to back it.

This isn’t about the scale of the problems the NHS deals with — the NHS just makes it visible. The lesson travels: investment doesn’t transform anything unless people move in tandem.

Hype aside, AI transformation raises the stakes, because it changes what work people do, and how they do it. For some, that evokes fear, anxiety, and uncertainty about what’s to come [2], alongside a real sense of loss. Loss of identity, autonomy, and the tasks that defined them [3].

The work itself will change, and that is inevitable, but the way we lead people can be the difference between a transition that moves or drags. With the change cycle becoming more frequent [4], any negative impact from a previous cycle only compounds in the next [5].

One train of thought is that change is situational, largely driven by external events, while transition is psychological — the process of an individual coming to terms with the situation that change has created [6]. Not acknowledging loss, or imposing direction without consultation, creates what Bridges calls a transition deficit — resistance, low morale, and disengagement. Bridges proposes three questions:

• What is changing?

• What will actually be different because of the change?

• Who’s going to lose what?

Although these provide a useful spine for planning, they are hardly script ready. Here are several I’ve used in the past, which not only support planning, but help you understand how emotions play a part in your change programme:

What did the old way give you that the new way might not? Surfaces loss without requiring people to name it as loss and gives you a sense of where their head is at.

Which parts of your role energise you? Helps surface where their motivation comes from. What energises someone tells you what they’ll fight to protect, and where the change is most likely to land badly.

How do you see your role changing, and why? Helps them start to imagine what the role could look like. Even if there’s no clear answer now, it’s a question worth coming back to. The response can also highlight whether more work is needed with their manager or an exec sponsor to clarify what the change means for them.

Where do you think your judgement still matters? Particularly useful in the AI context, where human judgement still works alongside the tech, especially in compliance-heavy environments.

What feels uncertain right now? Gives people permission to surface their fears in a safe space — just make sure it is a safe space.

These are all common sense, but seldom the first place people go when discussing a new product or process change. The default response is usually to lead with the benefits rather than spending time grounding the change in understanding first. I am not saying that the benefits aren’t important, but in my experience, they won’t land properly until people have identified what they’re losing and can figure out what that means for them.

On the flip side, there is a limit. All of this assumes people are rational — that if you surface the loss and give them room to name it, people will come round. Most will. Some won’t. There’s a difference between someone working through a transition and someone who has decided not to move, and no amount of careful questioning closes that gap. Knowing which one you’re dealing with is its own skill — and dealing with the second is a different beast entirely.

Sources

1. National Audit Office (2025). NHS England’s management of elective care transformation programmes. Session 2024-25, HC 707, 24 March 2025. nao.org.uk

2. Brougham, D., & Haar, J. (2018). Smart Technology, Artificial Intelligence, Robotics, and Algorithms (STARA): Employees’ perceptions of our future workplace. Journal of Management & Organization, 24(2), 239–257. doi.org/10.1017/jmo.2016.55

3. Selenko, E., Bankins, S., Shoss, M., Warburton, J., & Restubog, S. L. D. (2022). Artificial Intelligence and the Future of Work: A Functional-Identity Perspective. Current Directions in Psychological Science, 31(3), 272–279. doi.org/10.1177/09637214221091823

4. O Morain, C., & Aykens, P. (2023). Employees Are Losing Patience with Change Initiatives. Harvard Business Review, 9 May 2023. hbr.org

5. Bernerth, J. B., Walker, H. J., & Harris, S. G. (2011). Change fatigue: Development and initial validation of a new measure. Work & Stress, 25(4), 321–337. doi.org/10.1080/02678373.2011.634280

6. Bridges, W., & Bridges, S. (2017). Managing Transitions: Making the Most of Change (4th ed.). Nicholas Brealey Publishing.

When metrics move, it triggers an equal mix of curiosity and excitement. Not for the number itself, but for what it means. Something has changed.

The question I keep coming back to: how often do we build investment cases around customer research and internal objectives, without the broader market context?

In my experience, annual reports are one of the most underused sources of competitive intelligence available. They’re public, primary source, and directly from the business — signalling where competitors are investing, where they’re struggling, and where the risks and opportunities lie. Most product managers never open one, and I think that’s a missed opportunity.

To illustrate, I’ve pulled together an analysis of ASOS and Boohoo, extended from a study I conducted in 2022 to the latest year. All data taken directly from their investor sites.

Before we hit the rails

Before diving into metrics, a quick read of the industry helps identify which ones matter. Here’s what shapes this one:

• E-commerce is capital intensive — technology, infrastructure and fulfilment don’t come cheap

• Consumer loyalty is earned every transaction

• Suppliers and inflation squeeze from the cost side simultaneously

• Inventory management is where strategy meets reality

For simplicity, I’m focusing on a subset of metrics, although these should still help us understand some of the dynamics at work across these companies:

Asset utilisation — how much revenue the business generates for every £1 invested in assets: warehouses, technology, infrastructure, and working capital including stock.

Inventory days — how many days of stock a company holds, important for understanding stock turnover, tracking liabilities and avoiding stagnation. For online retailers, high returns rates can inflate this further as returned stock re-enters the warehouse.

ROCE — for every £1 invested in the business, how much operating profit is generated after costs. Positive means returns; negative means capital is being destroyed faster than it can be earned back.

Were they dressed for success?

In 2019, ASOS was generating £2.7bn in revenue, more than three times Boohoo’s £857m. Yet Boohoo generated £58.7m in operating profit against ASOS’s £35.1m. Scale wasn’t the advantage it appeared.

Boohoo was shifting stock twice as fast as ASOS — 63 days versus 140 days. Boohoo was a lean operator, built around small production batches, rapid stock replenishment and own-brand inventory. This was a key differentiator, providing a nimble response to new fashions without locking up significant capital.

ASOS had capital tied up in inventory for over 4.5 months. Although ASOS have been growing own brand products, which provides increased control, growing the number of partner retailers naturally stress tests inventory management across more diverse product lines.

The asset utilisation gap is worth noting. ASOS was generating £5.78 of revenue for every £1 of capital employed — nearly twice Boohoo’s 3.09x. This shows that the larger, more complex business was working its asset base harder. Both models were working, but in different ways.

Despite this complexity, both businesses were generating returns — Boohoo at 21% return on capital employed, ASOS at 7%. Two different models, both working until 2022–24 where returns on capital turned negative.

What did the acquisition actually cost?

Boohoo’s acquisition of the Arcadia brands and Debenhams represented a fundamental shift in scale and complexity. What had been a lean e-commerce business became a marketplace for a broader portfolio of brands. The consequences were significant — inventory days doubled after the acquisition, driven by taking on existing stock from new brands and dealing with increased supply chain complexity.

ROCE tells a similar story. In 2019, for every £1 invested, Boohoo generated 21p in profit after operating costs were deducted. By 2022–24, that had turned negative — the acquisition bets, the new distribution centres, and the expanded brand portfolio were consuming capital faster than the business could generate returns on it.

ASOS, which had generated 7p of profit for every £1 invested in 2019, faced its own version of the same pressures. High stock levels, a reduction in demand and US distribution investments that didn’t pay off resulted in an 11p loss for every £1 spent by 2022–24. For a business of ASOS’s scale, that amounted to roughly £220m of operating profit destroyed per year.

Both businesses were caught between the same structural forces — price-sensitive consumers, supplier cost pressures, inflationary headwinds, and an increasingly competitive market where companies like Shein were establishing themselves at exactly the price point both businesses had built their models around.

Are they showing signs of life?

In 2025, both businesses show signs of rebounding, holding less stock, reducing liabilities and running a more efficient operation. Boohoo, now rebranded as Debenhams Group, has moved toward a diversified marketplace model. ASOS has invested in reducing inventory cycles.

However, the overall picture is more cautionary. Operational efficiency is improving, inventory days are falling for both businesses, asset utilisation is recovering, while ROCE remains deeply negative for both. The operating machine is running better, but the returns have not yet followed.

What’s the bottom line?

Both businesses are significantly smaller than they were at their peak. Boohoo’s revenue has fallen from £1.98bn in 2022 to £790m in 2025. ASOS from £3.94bn to £2.48bn. The revenue decline isn’t simply a market story — the annual reports tell us why. Active customer numbers have fallen sharply for both businesses. For ASOS, active customers dropped from 26.4m at peak to around 18m by 2025. The revenue contraction follows directly from that customer loss. The operating machine may be improving, but the audience it is serving has contracted significantly.

That makes the path back to positive ROCE considerably harder than the operational metrics alone suggest. The revenue base on which returns need to be rebuilt has already contracted sharply — and rebuilding it requires winning customers back in markets that are more saturated than they were in 2019.

Understanding your market and where to differentiate

Annual reports surface the structural dynamics shaping a market — where competitors are investing, where they are struggling, and where the operating model is under stress. In fast fashion, inventory days is one of the most revealing signals. How much stock a business holds, how fast it moves, and how well supply chain processes support a nimble response to changing trends — these are not just operational details. They are strategic choices with direct consequences for capital efficiency and competitive positioning. Understanding where a competitor is exposed creates the space to differentiate.

Making smarter investment decisions

For product managers, this context sharpens rather than dominates the investment case. The decisions we make — what to build, where to invest, and how to prioritise — need to be grounded in more than customer research and internal objectives. Understanding the market dynamics, the value chain, and where competitors are generating or destroying returns gives you more intelligence for deciding where to place your bets. That rigour matters — not just for making better decisions, but for asking incisive questions.

Business acumen as a product skill

There is rarely more than one degree of separation between the product decisions a team makes and how those decisions land in the metrics that matter to the business. Customer retention shows up in active customer numbers, average revenue per customer and order frequency. Feature investment shows up in asset utilisation. The quality of product decisions shows up, eventually, in ROCE.

For public companies those signals are visible and accountable in the annual report. The same logic applies wherever capital is being invested in building products — the signals are just harder to see when they’re not publicly visible.

The metrics I’ve used here — asset utilisation, inventory days and ROCE — are three of many that can be derived from a standard annual report. They’re not complex to calculate. What takes judgement is knowing which ones matter for your industry, what they reveal about the operating model beneath the surface, and how to connect the financial signal to the strategic question you’re trying to answer.

That’s the skill worth building.

All figures used in this analysis were taken directly from the Boohoo Group plc and ASOS plc Annual Reports 2019–2025, available from their respective investor relations sites. 2020–21 figures have been excluded from period averages due to pandemic distortion of operating conditions. The 2022–24 figures are shown as a three-year average to smooth single-year volatility. Boohoo FY2025 reflects continuing operations only following significant changes to their portfolio.

Hello, I’m Terry.

I’ve been building and launching products since the early 2000s. Much of that time has been spent working with teams on the reality of transformation - not just the strategy, but the execution.

The working group where stakeholders pull in different directions. The pilot that doesn’t deliver what the spreadsheet suggested. The executive review where progress sounds better than it feels.

With technology continuing to accelerate, competitive advantages eroding faster than they once did, and organisations are under constant pressure to adapt - not once, but repeatedly. Some manage to keep changing. Others revert.

I’m interested in the quieter forces behind that difference: incentives, bias, ownership, ambition, fatigue, and what makes change actually hold.

How to actually bring people along. The advice I give product managers. How to build the conditions that help performance last.

This is where I’ll explore those questions - informed by research, tested in practice, discuss where both fall short.

You can expect one short post, early morning, mid-week. Reflections from practice, research that sharpens thinking, practical lessons from running tense sessions, preparing for senior scrutiny, and spotting early signs that performance may drift.

We’ve all been there. You’ve delivered, the numbers back it up, and yet one question quietly follows you out of the room: what’s going to stop performance from drifting?

It’s not pessimism. It’s a pattern. The same forces that drive markets toward equilibrium operate inside organisations — quietly, structurally, and often invisibly until it’s too late to course correct.

Like financial markets, complex biases create periods of overconfidence followed by correction. Organisations are no different.

No matter how confident leaders are about their own and their team’s ability to deliver transformative change, there are hidden forces at work. Forces that quietly pull what could be stellar performance back toward the mean.

There will be peaks and troughs. But in most cases, most of the time, firms revert to average performance over the long run.

That can sound melodramatic. Yet there is a real tension between what firms promise to boards, to investors, to themselves, and what transformation initiatives actually sustain.

Why performance erodes

Most of the complexity sits in structural forces — not in whether the change programme had a good launch deck, or a well-run steering group, although that all helps in the short term. It goes deeper.

As superior performance is sustained, high margins attract competition, eventually compressing profits. Companies that meet aspirational targets are more likely to loosen them and accept reduced performance. This is not through complacency or laziness, but a response to success and perceived risk. The urgency to seek and solve problems or find competitive advantage also lowers, leading to erosion in the long term.

Common failure modes include:

• Monitoring weakens after early success

• Targets soften; variance becomes accepted

• Incentives loosen their connection to results

• Competitors imitate what worked

• Costs creep back as complexity increases

What Intel did differently

Andy Grove recognised this dynamic at Intel. He didn’t treat strategy as pure intent — he operationalised it:

• Aligned incentives around the future of the business, not its past

• Abandoned profitable products that no longer fit the vision

• Created urgency around competitive inflection points

• Reinforced fast feedback loops

• Institutionalised measurable progress

This was the operationalisation of Moore’s Law — not as a slogan, but as a measurable discipline. Performance wasn’t left to drift. It was structurally reinforced.

The research is clear

Research across 732 firms shows the same pattern (Bloom & Van Reenen, 2007): companies that institutionalise measurement, target discipline, and incentive alignment outperform. Over 18 years of longitudinal studies from The World Management Survey supports these findings. There is no mystery. The difference is structural.

Performance frontiers

Grove used doubling transistor density and reducing the cost per unit to sustain performance and competitive advantage. This enabled Intel to continually reset the targets of computational power, even after success was achieved.

Each industry will have their own mechanical measures, each signalling how firms perform across a variety of competitive dimensions. A mechanical measure links user value to economic input — the tension between two metrics. It improves predictability, resists gaming, and supports alignment between targets and operations.

A few examples:

• Retail — inventory turnover × margin stability: users experience fresh stock while inventory remains profitable and disciplined

• Education — cost per completed learner × progression integrity: costs remain controlled relative to successful student outcomes

• SaaS — time-to-value × cost per active user: fast onboarding drives activation while cost per active user remains efficient

• Healthcare — referral-to-treatment time × cost per risk-adjusted outcome: patients are seen quickly while cost per accurate outcome remains managed

The leaders who sustain it

There are other factors that resist regression — chiefly, how effectively leaders manage and develop their people. Even with that covered, without relentless focus and structural congruence, the pull of gravity is difficult to resist.

The leaders I’ve seen sustain performance are the ones who treated structure as seriously as strategy — who asked uncomfortable questions even after success, and built the mechanisms to keep asking them. That discipline is rarer than it should be. But it is learnable.

Consider this in your own context:

• What metric defines your performance frontier?

• Has that frontier moved in the last 12 months?

• Did success raise or lower ambition?

• Would compensation change if that metric failed to improve?

• How quickly would we detect drift?

Further reading: Bloom & Van Reenen — Measuring and Explaining Management Practices Across Firms and Countries; Greve — Performance, Aspirations, and Risky Organisational Change; Grove — Only the Paranoid Survive.

I recently attended an MBA module at Warwick University on behavioural finance. A group of around 20 of us explored how biases, incentives, agendas, and emotions influence decision-making — both our own and that of others.

A recurring theme was how these signals fuel herd mentality.

That, in turn, can drive overinflated stock valuations, euphoria, and sustained bullish markets.

A lot of parallels were drawn between what we’re seeing now with AI and the dot-com bubble. Not as a perfect analogy, but as a useful reference point.

There’s already plenty of speculation about AI in the media. But working through behavioural biases in detail made the discussion feel more grounded — and more uncomfortable.

It reinforced the idea that we may be in a period where AI’s impact is being priced aggressively.

That led to a lot of debate in the room:

• Is the AI bubble similar to dot-com’s boom-and-bust?

• If we’re following a “bubble-like” trajectory, what would cause it to burst?

• How do investors and firms reduce downside risk?

The technology is real — the question is the valuation

There’s no doubt that AI is transformative. Like the internet, it represents a genuine step-change in how data and decisions are used.

Investors can see that promise. Today, a significant share of venture funding is flowing into AI enablers, platforms, and adopters. Even a private company such as OpenAI — arguably a late-stage start-up — is roughly valued at $500bn, with annualised revenue of around $20bn. That implies a valuation multiple of roughly 25× annual revenue.

This is where perception and reality begin to diverge.

Reading the price-to-sales signal

The divergence is most visible when looking at price-to-sales ratios — a rough signal of how much future growth is already being priced in.

By early 2026, the S&P 500 was trading at a trailing price-to-sales multiple of roughly 3.4–3.5, close to historical highs.

Against that backdrop, companies at the centre of the AI narrative stand out. NVIDIA, with a market capitalisation of roughly $4.4–$4.6 trillion and around $130bn in annual revenue, is effectively being valued at about $35 of market value for every $1 of current annual revenue. That expectation implies extraordinary future growth.

Even Apple, trading at a price-to-sales multiple well above the index average, reflects similarly optimistic expectations.

None of this proves valuations are wrong. But it does highlight how much of AI’s anticipated impact is already embedded in prices, and how sensitive those expectations may be to shifts in sentiment or realised returns.

Concentration risk

When many of the top 10 companies depend on the same resources — semiconductors — the diversification benefit of the index is weaker than it appears.

NVIDIA now accounts for roughly 7.4% of the S&P 500, concentrating that dependency further.

The gap between expectation and demonstrated impact is where things get more interesting.

The correction question

As more companies exploit AI to increase productivity and differentiate their offerings, strategies increasingly cater to market trends, with investors following these signals.

When the dot-com bubble burst, the market corrected and a more cautious approach followed. In the long run, the technology still transformed the world.

However, if firms don’t deliver on expectations, and if investors begin to substitute market signals with a more rational stance, is it only a matter of time before we see a similar correction?

That remains to be seen.

What’s clear is that the same behavioural forces we studied in that seminar room — overconfidence, herd mentality, narrative bias — are visible in the market right now. Recognising them doesn’t tell you when the correction comes, or whether it does at all. But it does change how you think about the decisions you’re making in the meantime.

Data sources: Multpl.com. Price-to-sales figures reflect early 2026 market data. Valuations are approximate and subject to change.