पाठशाला Pathshala · वृद्धि Vṛddhi, Growth · Lesson 26 · Scale

Brand when performance marketing plateaus

When every extra rupee on Meta and Google buys fewer customers than the last, the answer is rarely a better campaign. Decide when to fund brand, how to measure it honestly and the mix that keeps CAC flat.

Pathshala, The Founder Library · 11 October 2026 · 7 min read

Photograph: Rino Adamo · Pexels

The performance dashboard still shows a return above one. It has shown a lower one every month for six months, and each new campaign, creative and audience buys a week of relief before the line bends back. That is not a campaign problem. It is the market telling you that everyone who was already looking has been found.

Performance marketing, the paid search, social and marketplace ads that are bought and judged on clicks and conversions, is the right engine for a young company: it finds the buyers who are ready now and charges per result. The [paid acquisition lesson](/library/paid-acquisition-meta-google-and-the-ceiling) explains the ceiling it hits. This lesson is what comes after the ceiling: what brand does that performance cannot, when to start paying for it, how much, and how to know whether it is working without lying to yourself in either direction.

What a plateau looks like in the numbers

A plateau has a signature, and it is worth confirming before moving a rupee. Paid CAC rises as spend rises, while conversion on your own site holds steady: the problem is in the auction, not the product. The share of new customers who arrive through branded search, direct visits or word of mouth has stopped growing. Frequency on social campaigns climbs, meaning the same people see the ad more often. New audiences and creatives work for a week and decay. And the [channel-level CAC](/library/blended-cac-lies-channel-level-truth), not the blended one, shows the marginal rupee buying far less than the average rupee.

If conversion on the site or in the sales process is falling instead, the plateau is a product, pricing or positioning problem, and brand money will not fix it. If retention is weak, every channel will plateau early, because the [leaky bucket](/library/retention-is-the-growth-engine) needs ever more customers to stay level. Rule both out first.

Why most of the market is not buying this month

The clearest statement of why performance runs out comes from business buying. Peter Weinberg and Jon Lombardo of LinkedIn’s B2B Institute, writing in Marketing Week in September 2021, set out what they call the 95:5 rule, from research with John Dawes of the Ehrenberg-Bass Institute: at any time only about 5 per cent of business buyers are in the market, and the other 95 per cent will not buy for months or years. Their worked example is business banking: if companies switch banks about once every five years, about 20 per cent are in the market in a year, about 2 per cent in a month and about 1 per cent in two weeks.

Almost nobody on this promenade is shopping for your category today. Brand is how they remember you on the day they are. Photograph: Fahad Puthawala · Pexels

Performance marketing can only reach the few who are buying now, and every competitor bids for the same few. Brand advertising reaches the many who are not buying yet, so that when their moment comes your name is the one they already know. Consumer categories differ in how often people buy, but the shape is the same: the buyers ready this week are a small and contested pool, and the pool of future buyers is large and cheap to reach before the auction starts.

What brand does that performance cannot

Brand works through memory, and memory takes time to build and time to fade. It makes a buyer search for your name instead of the category, which takes them out of the auction. It makes every performance ad convert better, because a known name is clicked and trusted more. And it brings customers who never click an ad at all.

Airbnb is the public example of a company that learnt how much of its demand came from that memory. Its prospectus of November 2020 reported that about 91 per cent of all traffic in the nine months to September 2020 came through direct or unpaid channels, in a year in which, as part of its pandemic cost cuts, it had suspended substantially all discretionary marketing spend; it reported bookings beginning to recover by May 2020. Its first annual report then listed the reduction in performance marketing to focus on brand marketing among its plans, and described a model that let it be less reliant on performance marketing. Few companies have that much stored memory, and none builds it in a quarter. That is the point: the companies that can cut paid spend without collapsing are the ones that built brand before they needed it.

How much, and the lag

Les Binet and Peter Field’s work for the IPA, The Long and the Short of It, is the reference for the trade-off between short-term activation and long-term brand-building. Their later research with LinkedIn on business marketing, reported by the B2B Institute in October 2019, put the balance for B2B at about 46 per cent of budget on brand-building and 54 per cent on activation. Treat these as the direction of a mature mix, not as a target for next month. A company spending everything on performance should move in steps, a tenth of the budget at a time, and measure each step.

The step will hurt before it helps. Money moved out of performance stops buying customers at once; the brand it buys builds over months. What makes the move worth it is two effects working together: the paid rupees that remain buy more cheaply, because the most expensive marginal spend was cut first, and the brand lift arrives on every channel once built. The figure lets you set both and see when, or whether, the mix overtakes an all-paid budget.

What the brand money buys matters as much as how much. Brand work reaches people who are not looking, so it lives in broad media: online video, connected television, outdoor in the cities that matter, sponsorships and content that the target buyer meets in the course of a week. It works through the same few things repeated: one clear idea of what the company is for, the same visual and sound cues every time, and enough weight in a market for long enough to be remembered. Spreading a brand budget thinly across every city and every format for a month buys almost nothing. Concentrating it on two or three cities for two quarters buys something you can measure.

At the defaults, a ₹50 lakh monthly budget with a ₹2,500 paid CAC buys 2,000 customers a month. Moving 30 per cent to brand costs about 220 customers in month one. If the lift is 8 per cent per tenth of budget, the mix overtakes the all-paid line in month nine and ends two years about 420 customers ahead, with blended CAC near ₹2,350 against ₹2,500. Set the lift to 4 per cent and it never catches up: the two years end about 3,000 customers behind. The whole decision rests on the lift, which is why the next section matters more than the split.

Performance harvests the buyers who are looking. Brand decides whose name they look for.

Measuring brand without fooling yourself

Brand cannot be measured by clicks, and that is no excuse for not measuring it. Use four instruments together. Branded search volume and direct traffic, by month and by city: the cheapest leading indicator, and the first to move. Awareness surveys: ask a sample of your target buyers each quarter which brands in the category they can name unprompted and which they would consider, with the same questions every time. Geographic holdouts: run the brand campaign in some cities or states and not in matched others for a quarter, and compare new customers, branded search and paid conversion; the [attribution lesson](/library/attribution-what-actually-drove-the-sale) covers the design. A marketing mix model once you have two years of weekly data by channel: Google’s open-source Meridian lets a company run its own model in-house, with geographic data and reach and frequency where available.

Then judge brand on what it should change: branded search and direct share up, paid conversion up in the cities that saw the campaign, unaided recall up among the target buyers, and blended CAC flat or falling over four quarters while spend rises. If none of these moves in two quarters, the creative or the media is wrong, and more money will not fix it.

The monthly mix review

Once a month, the founder, the head of marketing and whoever owns finance spend an hour on one page. Paid CAC by channel against the month before and against spend. Branded search, direct traffic and their share of new customers. Paid conversion in the brand cities against the holdout cities. Blended CAC, with the trend over six months. Once a quarter, add the awareness survey and decide the next step in the split: move another tenth to brand if the leading indicators rose, hold if they were flat, and change the creative before the budget if they fell. Write the decision down with the number that drove it, so the next review can check it.

The figure is a model with your assumptions in it; the evidence for any lift is your own holdout. Move a tenth of the budget, not a half.