Audience & customer research: know who you're actually selling to

Most small businesses design for who they imagine their customer is. Customer research shows you who they actually are, built from interviews with your real buyers, not assumptions in a workshop. We turn that into three things you'll use every week: a persona, an ideal customer profile, and a journey map. The gap between the imagined customer and the real one is where budget gets wasted, and closing it is the whole point.

We're a small UK business in Peterborough, doing this for other small businesses. So we'll say the quiet part out loud: research costs time and money, and it doesn't guarantee a sale. What it does is replace guesswork with evidence, so the next decision (the website, the ad, the email, the product page) is made for the person who actually buys, not the one you hoped for.

Who this is for

This page is for UK small businesses — e-commerce stores on Shopify or WooCommerce, and B2B service firms — who recognise at least one of these:

You're launching or relaunching a website and want it built for real buyers, not a guess.
You're running ads that aren't converting, and you're not sure who you're actually talking to.
You sell to several different customer types but treat them all the same.
You're producing content that doesn't land, and you can't say precisely why.
You keep losing deals to competitors you can't clearly differentiate yourself from.

If any of those sound familiar, the problem usually isn't the website, the ad or the content. It's that the picture of the customer underneath all of them was never checked against reality.

The case for checking it is not subtle. Across the UK, only 21% of businesses that hold digitised data actually analyse it to generate new insight, according to the UK Government's UK Business Data Survey 2024 (DSIT/Ipsos, fieldwork October 2023 to February 2024). For smaller firms the figure is lower still: 34% of small businesses, 22% of micro businesses, and 20% of sole traders. For the overwhelming majority of smaller firms, the baseline is guesswork, which means a modest amount of real research goes a long way.

What you get

Three named artefacts. Not slide-deck filler, not a document that gets pinned to a wall and forgotten. Tools you point to when a decision comes up.

An ideal customer profile (ICP)

The firmographic, environmental and behavioural attributes of the accounts or buyers most likely to become your most valuable customers. Gartner's framework defines the ICP as exactly this, developed through qualitative and quantitative analysis of your historical customer data: your CRM, your order history, your reviews. It becomes the filter you qualify new leads and campaigns against: who to chase, who to politely decline.

Research-grounded personas

A persona is a fictional but realistic description of a typical or target customer, grounded in actual research, not a demographic stereotype. The Nielsen Norman Group's reasoning for why they work is practical: people remember a concrete person far better than an abstract statistic, so a persona becomes shared shorthand in every meeting. Instead of arguing about what "the customer" wants, you point to the research. A good persona also helps you resist the trap of designing for everyone, which means designing for no one.

A customer journey map

A visualisation of the process a person goes through to accomplish a goal, enriched with their actions, mindsets and emotions across every touchpoint. That's the Nielsen Norman Group's definition, and it's the one we work to. It shows where customers drop off, hesitate, or are delighted: the things your analytics record but can't explain. In a Nielsen Norman survey of more than 300 UX practitioners, journey mapping was rated most effective at educating teams about user pain points (4.06 out of 5) and at creating internal alignment (3.97 out of 5), and "involving customers in the process" was the single top-ranked success factor, which is exactly how we run it.

Optionally, a Jobs-to-be-Done frame. Where it helps, we'll add Jobs-to-be-Done statements that reframe your product from a list of features to the outcome the customer "hires" it for. The Nielsen Norman Group treats this as complementary to personas, not a replacement: personas build empathy and tell apart your customer types; Jobs-to-be-Done keeps the focus on outcomes. Useful when you need to think about what customers are trying to achieve rather than who they are.

Each of these feeds straight into the rest of the work: content strategy, SEO and search-demand priorities, website information architecture, campaign targeting, and email sequences. Research here isn't a standalone exercise. It's the foundation the later decisions stand on.

How we do it

We treat this as the discovery phase: researching the problem space and gathering evidence before anyone designs a solution. The Nielsen Norman Group frames discovery the same way, and lists personas, journey maps and user-needs statements as its natural outputs. Here's the shape of the engagement.

1. We start from what you already have

Your CRM, your analytics, your reviews, your sales notes, and a quick competitor and benchmark review. Plenty of evidence usually exists already. It's just never been read as a whole.

2. We interview your real customers

Depth interviews are an attitudinal, qualitative method: they surface the pain points, mental models and motivations behind what people do. We interview your actual buyers, not a panel of strangers.

The key limitation, which the Nielsen Norman Group is clear about, is self-reporting bias: what people say they do isn't always what they do. That's why we validate interview findings against your real behavioural data, and where it matters add field observation or a short diary study to check what people say against what they actually do.

3. We recruit to saturation, not to an arbitrary number

A common worry is "how many people do we have to interview?" The answer is evidence-led, not a round figure. Research the Nielsen Norman Group cites found that 20 to 30 interviews uncover 90–95% of all customer needs (Griffin & Hauser, 1993), and a separate 60-interview study reached saturation at 12, with the high-level themes already clear after just 6 (Guest, Bunce & Johnson, 2006). For most small businesses with a focused question, that means we start with five or six, analyse as we go, and keep recruiting only until new interviews stop telling us anything new.

It's worth not confusing this with usability testing, where five users is famously enough. The Nielsen Norman Group's analysis of 83 of its own consulting projects found testing more people barely added insight. Discovery interviews aren't usability tests; they typically need a larger, iterative sample. We scope the number to your question, not to a template.

4. We involve your team throughout

Personas and journey maps fail in predictable ways, and the Nielsen Norman Group has documented them: when they're built in isolation by an outside team and unveiled like a piece of artwork; when they have no specific, well-defined goal; and when nobody ever actually uses them in a decision. We avoid all three by keeping your team in the loop from the first interview to the final artefact, and by tying every output to a real decision you're about to make.

A note on data protection

Interviewing and surveying customers means processing personal data, so it's fair to ask whether there's a UK GDPR problem. There is one to manage, and it's manageable.

For private-sector market research, the most likely lawful basis is legitimate interests, subject to the Information Commissioner's Office three-part test of purpose, necessity and balancing. The Data (Use and Access) Act 2025 introduced a separate "recognised legitimate interest" basis, but that's reserved for five specific public-interest situations in its Annex 1 — things like safeguarding and responding to emergencies — and it does not cover ordinary commercial customer research. So standard legitimate interests, with the balancing test, is what applies. If we recruit interviewees by email or other electronic outreach, PECR applies to that too.

We run research in line with ICO guidance and advise you on data handling as part of the engagement. Most small businesses don't realise they need to think about this, so we do it for you.

What this isn't

Plain about the things that get sold as "customer research" but aren't:

Not proto-personas from a two-hour internal workshop. Personas built on team assumptions with no customer contact are, in the Nielsen Norman Group's own terms, only suitable for teams who would otherwise skip research entirely. They confirm what you already believe instead of testing it.
Not AI-generated buyer profiles. AI can produce plausible-sounding demographic descriptions, but plausible isn't true, and an AI persona carries no evidence that it reflects anyone who actually buys from you.
Not demographic stereotypes, and not a static document destined to be filed and never opened again.

Research-based personas change decisions. Assumed ones just dress up the decision you'd already made.

Why True Noise

We're a small business in Peterborough serving small businesses across Cambridgeshire and the UK. That matters in two ways. First, we understand your context from the inside: the budget that can't absorb repeated mistakes, the need for one or two clear answers rather than a research department. Second, local market knowledge is part of the work: a B2B buyer in Cambridgeshire distribution, agri-food, engineering or professional services doesn't look like a London-centric template, and a researcher who knows the local ground reads the evidence better.

Research-first is simply how we start every project: strategy, then research, then build, then optimise. Customer research isn't an upsell bolted onto a website; it's the default first step, because building on a guess is the expensive way to do it.

Our position rests on the evidence above — the research methodology drawn from the Nielsen Norman Group, Gartner's ICP framework, the saturation studies (Griffin & Hauser, 1993; Guest, Bunce & Johnson, 2006), and the buying-gap figures from Forrester and the UK Business Data Survey 2024.

What it costs

We scope every engagement to the size of the business and the stakes of the specific decision it needs to make. A single well-grounded persona tied to one decision costs very differently from a full ICP, persona set and journey map ahead of a relaunch.

Research and discovery work is quoted as a custom project (our project rate is £60 per hour, with a scoped proposal first), and it slots neatly into our monthly plans when you want it to feed continuous content, SEO and campaign work. You'll get a clear scope and price before anything starts. See our pricing for the full picture, or get in touch and we'll scope it with you.

FAQ

Do we really need to pay for research? We already know our customers

It's the most common objection, and the evidence runs against it. The Nielsen Norman Group finds that personas fail precisely when a team is sure it "already knows" its users. And Forrester's State of Business Buying 2024 (4 December 2024) found that 81% of B2B buyers are dissatisfied with the provider they chose, which suggests the gap between what sellers think buyers need and what buyers actually need is systematic, not the odd unlucky deal. Research replaces belief with evidence. If you're right about your customers, it confirms it cheaply; if you're not, it saves you far more than it costs.

How many customers do you need to interview?

Enough to stop learning new things, not a fixed quota. For qualitative persona work, saturation is typically reached around 20–30 interviews (Griffin & Hauser, 1993, cited by the Nielsen Norman Group), but a focused question with a narrow audience can be answered with as few as 5–12. We start with five or six, analyse as we go, and recruit to saturation. The aim is insight, not statistical significance.

What's the difference between an ICP and a persona?

An ICP is account-level: the firmographic, environmental and behavioural attributes of the accounts or buyers most likely to become your most valuable customers (Gartner). You use it to decide who to target and qualify. A persona is people-level: a research-grounded, narrative picture of a real customer segment (Nielsen Norman Group). You use it to decide how to communicate and design for them. For B2B clients, both earn their keep — the ICP guides who, the personas guide how.

Can't AI just generate personas for us?

It can generate something that reads like a persona. The trouble is that a persona built without real interviews is a proto-persona — a shortcut built on assumptions — and the Nielsen Norman Group is explicit that proto-personas suit only teams with very low research maturity who would otherwise do nothing. An AI-generated profile carries no evidence that it reflects anyone who actually buys from you. We'd rather interview six real customers than invent sixty plausible ones.

What's a customer journey map and why would we want one?

It's a visualisation of the process a customer goes through to reach a goal, with their actions, mindsets and emotions mapped at each stage (Nielsen Norman Group). It reveals where people drop off, hesitate or are delighted: the "why" behind numbers your analytics can only show you as a drop. For a small business with several touchpoints — social, email, website, in person — it shows how those connect from the customer's side, not yours.

Yes, and it's manageable. Interviews and surveys process personal data, so you need a lawful basis. For private-sector market research that's most likely legitimate interests, subject to the ICO's three-part test (purpose, necessity, balancing). The Data (Use and Access) Act 2025's separate "recognised legitimate interest" basis is reserved for five public-interest situations and doesn't cover ordinary commercial research, so the standard balancing test still applies. We handle this in line with ICO guidance, including any PECR obligations if we recruit interviewees electronically.

How long does it take?

It depends on the scope: the number of interviews, how hard your customers are to reach, and how many artefacts you need. We'll give you a realistic range when we scope the work. We won't quote you a fixed timeline we can't stand behind.

We're a very small business. Is this worth it for us?

Arguably more so than for a large one — there's no budget to absorb the same mistake twice. The Nielsen Norman Group's principle is that persona scope should match the design challenge, built for your specific context rather than to a fixed number. We apply that directly: instead of a suite of statistical personas, a small business usually needs one or two well-grounded qualitative personas tied to a specific decision. That calibration is ours, made for your size and your stakes.

Start with the decision you're trying to make

Tell us what you're about to build, change or spend on, and we'll scope research to fit it — small and focused, or full discovery ahead of a relaunch.

Get a free audit — a quick, no-obligation look at where assumptions might be costing you, and whether research is worth it for your situation.
Start a project — scope the research with us, with a clear plan and price before anything begins.