One idea kept resurfacing at the Goldman Sachs 10,000 Small Businesses UK Summit: How Britain Can Win early in October 2026.
As businesses grow, it becomes harder to stay close to customers.
Not because leaders suddenly stop caring about them.
Quite the opposite. Growing organisations invest in more research, more data, more specialists, more systems and more sophisticated processes designed to help them understand and serve their markets.
Yet, somewhere along the way, it is possible to become better informed and less connected.
That matters because customer closeness is not simply a cultural virtue or a customer service issue.
It is a strategic capability.
It affects which markets you choose to serve, how you segment them, what you offer, where you invest and whether your proposition remains relevant as customer needs change.
The paradox of scale
Sahar Hashemi, co-founder of Coffee Republic, provided perhaps the clearest illustration.
Coffee Republic’s early growth was intensely customer-led. The founders were close to the people buying the product and could see and hear what customers wanted.
As the company grew, raised money and became more professionalised, more layers appeared between the organisation and those customers.
Hashemi described external agencies producing increasingly sophisticated customer personas which, in her experience, bore less and less resemblance to the actual customer she knew.
The issue isn't that personas, research or external expertise are inherently wrong.
The problem comes when the representation of the customer begins to replace the customer.
It is easy to see how this happens.
A founder may initially hear customer complaints personally, watch how people use the product, speak to lost prospects and know why customers come back.
Later, that same founder may receive a monthly dashboard summarising those interactions.
The information has travelled further, been processed more extensively and become increasingly abstract.
The business has gained scale.
But it may have lost signal.
Customer closeness is not the same as collecting more data
This is an important distinction.
Most established businesses have no shortage of customer data.
They have NPS scores, CRM systems, web analytics, customer satisfaction surveys, sales reports, social listening, service data and research programmes.
The question is whether those things are helping the organisation make better market choices.
Ken Murphy's discussion of Tesco offered an interesting contrast. His description of the business suggested an organisation using detailed operational information to remain intensely interested in what customers are actually experiencing.
He shared that Tesco understand that to win and retain customers, availability, cancellations and what happens in an individual store matters.
Why?
Because this is where the brand promise actually happens.
Similarly, Sean Doyle talked about British Airways learning directly from the experiences of passengers requiring additional assistance and using that understanding to improve the service.
These are examples of customer insight doing useful work.
This is not information accumulated because a dashboard says it should be measured, but examples of data informed evidence being used to identify a problem, make a choice and improve the customer outcome.
The customers you hear from do not necessarily represent the market
There is another danger as businesses grow.
Senior teams can become disproportionately close to their largest, longest-standing and most engaged customers.
Understandably, these are often the customers invited to events, included in research and given access to senior leaders.
Their views matter.
But they are not necessarily representative of the entire market.
The customer who has just bought from you for the first time may see the experience very differently.
The prospect who chose a competitor may know something enormously valuable about your proposition.
And the customer who quietly stopped buying may tell you more about future competitiveness than the loyal customer who has been with you for a decade.
Bad news also has a habit of travelling less effectively through organisations.
Lost business, unsuccessful bids and declining relationships can feel like failures to be explained rather than evidence to be examined.
The result can be a dangerous distortion:
leadership hears disproportionately from the customers who stayed versus those that are available to win.
Research can suffer from the same bias. Existing customers are easier to reach than competitors’ customers, non-customers or emerging segments.
That may be perfectly adequate when the objective is market penetration.
It is much less useful when the organisation is looking for new markets, new segments or future sources of revenue growth.
Customer closeness therefore means getting deliberately close to a cross-section of the market, not simply the customers who are easiest to hear from.
The strategic danger of becoming internally rational
Rory Sutherland offered one of the most memorable provocations of the Summit:
“Don’t be so rational that you can’t get lucky.”
He also warned against corporate isomorphism — the tendency for businesses in the same market to use the same metrics, suppliers, logic and benchmarks, and then gradually become more alike.
His story about restaurateur Will Guidara brought that to life. Guidara wanted to build the best restaurant in the world. When his leadership team visited one of the world’s best restaurants, he was disappointed that they came back with a list of things they could copy.
The instinct is understandable.
Study the best. Benchmark it. Replicate what works.
But if everyone studies the same examples, tracks the same measures and applies the same logic, the result may be improvement without distinctiveness.
The same risk appears as organisations grow and decisions become increasingly internally focused. Internal questions tend to orient around:
What can our systems support?
How do our existing teams operate?
Which products do we already have?
What can we deliver most efficiently?
What did last year's numbers tell us?
All are reasonable questions.
But none of these begin with the market!
Over time an organisation can become extremely good at making internally logical decisions which are becoming progressively less relevant to its customers.
This is particularly dangerous in market strategy.
For example, businesses frequently define their markets according to the products and services they sell.
Customers rarely see the world in quite the same way.
Customers have problems to solve, outcomes to achieve and alternatives available to them.
The competitor to a product may therefore be another product, a different technology, an internal solution or simply doing nothing at all.
Staying close to customers helps businesses see these shifts earlier.
It changes the questions from:
“How do we sell more of this product?”
to:
“What problem is the customer trying to solve, and are we still one of the best ways of solving it?”
That is a much more useful strategic question and provides space for something that excessive benchmarking and rationality can squeeze out – the distinctive idea that competitors haven’t yet come up with.
Founder mentality is really a customer-proximity discipline
There was considerable discussion at the Summit about maintaining a “founder mentality” as businesses grow.
That phrase can sound as if the answer is somehow to preserve the personality of the founder forever.
We think there is a more useful interpretation.
Founder mentality is, in part, about proximity.
Founders are usually closer to customer problems because there is nowhere else for them to go.
They hear the customer complaint.
They see the lost sale.
They know which promise they cannot deliver.
They spot the unusual request.
They know when a customer starts behaving differently.
The real challenge for a scaling organisation is therefore not to behave nostalgically like a start-up.
It is to design ways of retaining that proximity when the founder can no longer personally provide it.
Greg Jackson's comments about Octopus Energy added another dimension to this.
As organisations scale, decision-making often becomes professionalised alongside everything else. More reports, more approval levels and more processes are introduced.
Some of those are of course, necessary.
But if every customer signal must travel upwards through multiple layers before somebody has the authority to respond, the organisation has created distance twice: distance from the customer and distance from the decision.
Customer closeness therefore requires more than insight.
It also requires agency.
The people closest to customers need sufficient information, authority and clarity to act on what they are learning.
Be careful not to research the customer out of the room
None of this is an argument against segmentation, personas or quantitative research.
Quite the opposite.
Good market strategy requires evidence.
But the purpose of those tools is to make differences between customers more visible so that better strategic choices can be made.
They should sharpen reality rather than sanitise it.
A useful segment tells us something meaningful about differences in customers' needs, behaviours, priorities, economics or buying processes.
A useful value proposition then answers three practical questions:
Who is this offer particularly relevant to and what problem are we solving?
Why is our solution different or better than the alternatives available?
Why should those customers believe us?
Those questions are difficult to answer well from inside a boardroom.
They require evidence from the market.
But take care! Averages need treating carefully.
A market-level result can conceal very different responses between segments. If half of one audience thinks a price is too high while another half would willingly pay more, an average conclusion that “the price is about right” is strategically almost useless.
Good research should reveal the differences between customer groups, not smooth them into a market average. Because effective strategy is developed at segment level, the evidence informing it should be too.
AI makes customer closeness more important, not less
AI unsurprisingly featured heavily throughout the Summit.
There is enormous potential.
Businesses can analyse customer information more quickly, personalise interactions at much lower cost and put much better information into the hands of employees.
But there is an interesting tension here.
The technology that makes it easier than ever to know more about customers also makes it possible to become strangely remote from them.
James Quincey raised the question of where hyper-personalisation ends and the customer starts to feel “stalked”.
Ken Murphy emphasised the importance of building and maintaining trust.
Those are market questions, not simply technology questions.
Just because an organisation can personalise something does not mean the customer will value that personalisation.
The test remains the same:
Does this make the business more useful to the customer?
Does it solve a meaningful problem?
Does it strengthen or diminish trust?
AI absolutely can help organisations interpret the customer signal but it should not become another layer between the business and the people it serves.
Scale the business. Protect the signal.
Perhaps the most useful conclusion from the day is that customer closeness should not be left to chance.
As organisations grow, they need to design for it deliberately.
That means creating regular direct exposure to customers at senior levels — and including new, lapsed and lost customers, not merely the largest or most loyal.
It means ensuring service teams, sales teams and frontline employees are treated as sources of market intelligence rather than simply delivery functions.
It means testing segmentation and propositions against observed behaviour rather than relying solely on internally created assumptions.
And it means asking whether the metrics reaching the board are genuinely helping leaders to understand where customer needs, behaviours and alternatives are changing.
Growing businesses inevitably need more structure.
The challenge is to make sure the structure helps the organisation hear the signals from the market rather than muffling them.
Because the risk of scale isn't simply bureaucracy.
It is becoming operationally excellent at delivering something customers increasingly value less.
Three actions to take now
1. Measure your distance from the customer.
Ask every member of the senior team when they last personally heard a customer conversation, observed a customer using the product or service, reviewed a lost sale or saw the buying decision first-hand.
2. Test the abstractions.
Take one important customer segment or persona and compare it with recent customer evidence. Speak to current customers, new customers, lapsed customers, sales teams and service teams. Ask whether the description is making customers clearer — or simply more convenient to discuss internally.
3. Put customer signal into strategic decisions.
Choose one significant growth decision currently being considered. Before discussing communications or execution, ask: What problem are we solving? For which customers? What alternatives do they have? What evidence tells us their needs or behaviour are changing?
Then use those answers to shape the strategy.