The Line You Take on Trust

AI has changed how customers find you and what marketing costs. It has also made the return on every pound easier to measure.

You are looking at last year’s accounts. Halfway down the profit and loss account is a line headed advertising and marketing. It says £30,000.

You know roughly where the money went. Some on Google Ads, some to a social media agency, some on a stand at a trade show. What you can’t say is what each of those pounds brought back.

That is normal. For decades, businesses of every size spent on marketing and took the return on trust. The biggest could afford to, because their size did much of the work.

At the end of September, Mark Abraham and four colleagues at Boston Consulting Group (BCG) published new research. It carries a blunt message for the chief executives of consumer goods companies: they can’t take marketing returns on trust any more.

The first piece of evidence is breakfast cereal. Between 2021 and 2025, small brands that moved quickly on changing tastes in cereals and granola bars gained 10.8 percentage points of market share. The established brands, with the bigger budgets, lost more than 12.

BCG wrote for large companies. I read it as an adviser to owner-managed businesses, and most of it applies to you directly. In a £1m business making £100,000 profit, £30,000 of marketing equals 30% of that profit. Give it the test you would give any other investment: what will it return, and when?

1. What has changed

In the era of mass media, a large budget bought reach and reach bought sales. Nobody needed to know which pound did the work. BCG explains why that arithmetic has stopped working, and its reasons apply to a business of any size.

Customers now come across a business in many more places before they buy. BCG counts more than 15 points of contact on the way to a purchase, up from about 5 a decade ago. Each one needs something said, shown or written. The same budget now must cover three times as many places.

Marketing material has become far cheaper to produce. BCG’s example is a personalised photo of a beauty product. Two years ago it cost $4 to make. Now it costs 4 cents.

AI will cut your production costs too, so decide now what happens to the saving. Bank it, and your costs fall. Spend it on the channel that repays fastest, and it pays for growth.

The return has become easier to see. BCG’s central point is that AI gives chief executives something they have long lacked. They can now see more clearly what marketing produces, and where the next pound would earn the most. For a small business, most of what you need is already in Xero, plus the answer to one question: how did you find us?

You are not late. In BCG’s annual survey of 300 marketing chiefs, nearly all said AI is changing how their whole function works. More than two-thirds had not moved beyond the basics. BCG names David Protein and Poppi as challenger brands growing quickly with small teams that used AI from the start.

2. Why marketing is the first cost to go

When profit falls, marketing is often the first cost an owner cuts. It is quick to stop, and its return is hard to see. BCG describes the same habit among chief executives: marketing costs rise, growth slows, and the budget gets cut. It calls the habit a reflex left over from an earlier era.

The timing makes that reflex expensive.

The saving appears in next month’s accounts. The lost sales arrive months later, and nobody links them to the cut.

I have argued for years that marketing and training deserve more investment in a downturn. That argument comes with a condition: you need to know which marketing is working. Cut blind, and you may stop the channel that was paying for itself. Measure first, and you can cut what isn’t working and keep the rest.

3. The next £1,000

Is there a better way? Start by splitting that single marketing line into the channels that make it up. Then count the new customers each channel produced.

Example: take the £30,000 from the opening, spread across three channels. Each new customer buys about £2,500 a year at a 40% gross margin. That gives £1,000 of gross profit a year, or about £83 a month, to repay the cost of winning them.

The average new customer cost £667 and repaid that cost from gross profit in 8 months. Inside that average is a channel where each customer costs four times as much as Google Ads and takes 2 years to repay.

Averages hide the channel that isn’t working.

Measure the return in gross profit, because gross profit pays for the marketing. A campaign that brings in £10,000 of sales at a 30% margin earns £3,000 of gross profit to cover its cost. If it cost £4,000, it lost £1,000. It needs repeat business to pay its way.

Then ask the question BCG puts to chief executives: where will the next pound earn the most? At your scale, make it the next £1,000. Returns tend to fall as spend in one place rises, so the next £1,000 on Google Ads may win fewer customers. Add spend in steps and watch the cost per customer as you go.

Before you cancel the social media agency, ask what job it does. BCG accepts that some marketing returns will always be hard to pin down. Spending that builds your name may help the other channels without producing customers you can trace.

Judge that spending over a longer period and by different measures, such as enquiries that mention your posts. Agree in advance what it should produce and by when. Then check.

4. When customers ask AI first

BCG surveyed 13,000 consumers in 12 markets. Nearly a third use AI somewhere on the way to a purchase, three times the share of 18 months earlier. About 20% use it routinely, and 13% end up buying what the AI recommends.

BCG also reports that search traffic through AI tools grew by nearly 150% in the year to July 2026. The survey covers consumers, and your customers may move at a different speed. I’d expect them to move in the same direction.

When someone asks an AI tool for a recommendation, two things happen. First, the tool decides whether to mention you, based on what it can find. Then it judges you, using everything it can read: your website, your reviews, and comments you didn’t write.

This applies even if most of your work comes by referral. BCG found that more than 40% of consumers aged 45 and over research and compare options before they buy. Among younger buyers the figure is more than half. A friend’s recommendation may now get checked with an AI tool before anyone picks up the phone.

BCG makes a further point. When a product falls short of what its marketing promised, AI tools can find the gap in reviews and conversations the brand doesn’t control. If your website promises a reply within 24 hours and three reviews mention waiting a week, a prospect may hear about both.

An AI tool reads what your customers wrote about you, as well as what you wrote about yourself.

Try this today. Ask ChatGPT or Claude the question a stranger would ask, using your own trade and town. For example: who is a good commercial electrician in Chester? Then ask what the tool knows about your business.

Compare the answers with how you would describe yourself. Where they are wrong or thin, the fix is usually plain. Make sure your website says what you do, for whom and where. Collect reviews that describe specific results in your customers’ own words.

If the reviews point to a gap between promise and delivery, close it in the way you work. Then repeat the test every quarter, because the answers change.

5. Write down how you win work

BCG describes a consumer goods business that built an AI tool to write campaign briefs from its raw data. At first, its briefs were only 50% accurate. The company then wrote down how its best marketers put a brief together, the checks they applied and how they answered the key questions. Accuracy rose to more than 90%, and the tool now saves the company’s 3,000-plus marketers weeks of work each year.

The know-how that wins your work sits mostly in your head. Who your best customers are. What they ask before they buy. What you say that persuades them. What you will never promise.

Write it down. A few pages will do. The BCG example shows how much better an AI tool performs once it has that kind of document.

Your know-how is the part of your marketing a competitor can’t easily copy.

Keep that document in a file you own. BCG warns businesses against tying their know-how so closely to one software supplier that switching becomes difficult. The same risk applies at your scale. If your marketing knowledge lives only inside one tool’s settings, it leaves when you change tools.

Keep the measurement too. BCG’s view is that agencies and technology partners can carry out the work while the measuring stays inside the business. Suppose your agency is paid a percentage of what you spend and also reports what the spending achieved. That arrangement creates a conflict of interest, however honest the agency.

BCG reports that agencies are moving away from fees based on a share of spend. Until yours does, check its reports against your own numbers in Xero.

6. What to do this month

Beyond the Return rests on three pillars. Management reporting that supports decisions while they still matter. KPI analysis that links behaviour to profit and cash. Structured advisory meetings that translate numbers into decisions. Marketing needs all three.

Start with the reporting. Set up a tracking category in Xero with one option for each marketing channel and code every marketing cost to it. Your monthly accounts will then show spend by channel within days of the month end. You can stop a weak campaign in its second month, while stopping it still saves money.

Then the KPIs. Ask every new customer how they found you, and whether they used an AI tool along the way. Record the answer against the customer.

Track four numbers each month. Enquiries by source. The share of enquiries that become customers. The cost of each new customer, by channel. The months of gross profit it takes to repay that cost.

Add one behaviour that drives those numbers, such as how quickly you reply to an enquiry, or how many happy customers you ask for a review.

Then the meeting. Give marketing 15 minutes at your monthly review and make one decision each time: where the next £1,000 goes. Any channel that can’t show a return goes on notice, with a named result and a date.

BCG gives a business 6 months to make its marketing scorecard part of daily work. You can have yours running by your next month end.

BCG closes with five questions for chief executives. Here are four for business owners to take into their next monthly review:

1. What did each new customer cost last quarter, channel by channel?

2. How many months of gross profit does it take to repay that cost?

3. Where will the next £1,000 go, and what result would make you stop?

4. What does an AI tool say about your business when a stranger asks?

7. What this means for your business

AI has made marketing material cheaper to produce. It has spread your customers’ attention across more places and given them a new way to judge you. It has also made the return on marketing easier to see, for a £1m business as much as for a global brand.

BCG told chief executives they can’t take marketing on trust any more. The same holds at your scale, and the work is smaller. A tracking category in Xero, one question for every new customer, four numbers and 15 minutes a month.

Know what each pound of marketing returns in gross profit and decide each month where the next £1,000 goes. When competitors cut blind, you can keep spending on what works.

Measure enough to decide. Then decide.

Sources and Further Reading

Mark Abraham, Jessica Apotheker, Yotam Ariav, Robert Derow and Lauren Wiener, “Why Most CEOs Are Failing Marketing’s New Math Test”, Boston Consulting Group, 29 September 2026, bcg.com/publications/2026/how-ai-rewrites-cpg-marketing-economics. The market share, survey and cost figures in this essay come from this article.

Noel Guilford FCA, “Beyond the Return: Second Edition”, Guilford Accounting, April 2026.

Noel Guilford, “Numbers Don’t Lie”, 2025.

About the Author

Noel Guilford FCA is a chartered accountant and business adviser who works with a small number of deeply engaged business owners through a structured monthly advisory programme, The Business Alignment Programme. He is the founder of Guilford Accounting and writes on practice design, advisory strategy, and the intersection of AI with professional judgement.

To discuss how this approach might work for your business, book a discovery call at calendly.com/noelguilford.