In October, Coca-Cola's global CMO stood up at the ANA Masters of Marketing and said: "With age and with experience, I do not believe any more in love."
Manolo Arroyo was talking about brand love, the perception metric Coca-Cola ran on from 2000 to 2020. It kept climbing while the business did not, and millions of people who said they loved Coca-Cola never drank any. The company replaced it with total consumer base, a behavioural metric. The line Arroyo added is the one worth keeping: "Love can be increased as well, but it follows the behavior."
Note what that is and is not. Coca-Cola did not stop building its brand. It changed what it counted: from what people say to what people do.
Which is the useful end of an argument that usually generates more heat than light. Ask a room of marketers whether brand or demand matters more and you will get an hour of theatre. Greg Boosin, CMO of Equitable, had the better answer at a Brand Innovators summit this autumn: "Don't take the bait and answer what's more important: brand vs. demand."
He is right, and there is a sharper version of the point. Demand work, done properly, is brand building. Not a substitute for it, not a lesser cousin of it. The same thing, arriving in a different order.
What the evidence actually says
Three findings worth holding together.
- Most of advertising's return shows up after the quarter you measured it in. Profit Ability 2 is a meta-analysis by Ebiquity, Gain Theory and others, covering 141 brands and £1.8 billion of media spend between 2021 and 2023. It found an average short-term profit return of £1.87 per pound. Measured across the full 24 months, it is £4.11. More than half of advertising's profit arrives after the window most performance dashboards close.
- Recruiting beats retaining. Binet and Field, working from 996 campaigns in the IPA Effectiveness Awards databank, found campaigns aimed at acquiring new customers roughly three times as effective on average as loyalty campaigns aimed at existing ones. Growth comes from more buyers, which is what demand generation is for.
- Share of voice drives long-term growth. From the same dataset: brands whose share of voice runs above their market share tend to grow, and brands below it tend to shrink. Buying attention at scale is not the opposite of building a brand. It is a mechanism for building one.
Put those together and demand-first stops looking like a shortcut. Work that reaches new category buyers at scale, and keeps reaching them, produces the effects the brand-building literature credits to brand building.
The condition
There is one, and it decides everything. Jean-Paul Jansen of Paula's Choice, at the same summit: "If your product is inferior, no amount of marketing can fix that."
Demand-first is a bet on the brand's ability to deliver on what the demand promised. Generate trial for something good and repeat purchase compounds: the demand you created becomes the brand you have. Generate trial for something that disappoints and you have bought an expensive education for customers who will not come back, and taught them something about you that the next campaign has to undo.
So the question to settle before the money moves is not whether brand or demand comes first. It is whether the product, the pack, the site, the shipping and the support can hold the people the campaign is about to send.
Where demand-first actually goes wrong
Not in choosing demand. In confusing two different things that sit in the same column of the spreadsheet.
- Creating demand means reaching people who were not in the market, or did not know the category could solve their problem.
- Harvesting demand means collecting intent that already exists: branded search, retargeting, recent-purchaser audiences, discount-led reactivation.
Harvesting reports beautifully. It has the shortest path to a conversion and the best-looking return in the deck. It also builds nothing, because the demand was already there. You paid a toll to collect a sale you had largely already earned. A programme that is mostly harvest looks efficient for about four quarters and then stops growing. That is the audience ceiling, and it arrives on schedule.
Which is why the honest version of this argument is demand-first, not demand-only. The evidence on mix is not ambiguous: Binet and Field put the long-run optimum near 60% brand building and 40% sales activation, and the UK profit work found most brands both over-weighted toward short-term channels and underspending overall. The average brand in that dataset could roughly double media investment before each additional pound stopped returning a profit.
Mark Ritson, reading Arroyo's remarks, made the point that marketing overcorrected toward loyalty and love in the 1990s and will probably overcorrect toward penetration now, and that love will likely return within a decade as a secondary metric. That sounds right. The useful position is not to pick a side of a pendulum, but to measure behaviour and let the attitude data explain it.
What this looks like in practice
- Split the two lines in your own reporting. What share of last quarter's return came from new buyers, and what share from demand you would have captured anyway? Most teams have never run that cut.
- Set a floor for demand creation. A fixed share of budget aimed at people who do not know you, protected from in-quarter raids when the monthly number looks short.
- Measure the slow part on a slow clock. Say up front that some of the return lands at month 12 to 24, and agree with finance how it will be read, so it doesn't get cut for being slow.
- Check the delivery side before you scale. Repeat rate, review scores, fulfilment times, support load. Demand multiplies whatever the experience already is, in both directions.
- Hold both to one number, agreed before the spend.
Brand and demand were never two budgets competing for the same meeting. One creates the customer; the other makes sure that customer can find you and buy. Run them as one programme and the question of which comes first stops being interesting.
A note on the data: Profit Ability 2 and the related Growth Gap analysis were commissioned by Thinkbox, the UK commercial TV marketing body, and both conclude that brands should move budget toward television. We find the profit-timing and saturation findings useful and read the channel recommendations with the sponsor in mind. Binet and Field's databank draws on IPA award entrants, which skews toward campaigns that worked.
The Drum: Mark Ritson on Coca-Cola's top marketer falling out of love with brand love (Oct. 2026)
Brand Innovators: CMOs & CFOs aim to “speak the same language” (Oct. 2026)
Thinkbox: Profit Ability 2, the new business case for advertising (Ebiquity, EssenceMediacom, Gain Theory, Mindshare, Wavemaker, 2024)
Thinkbox: Brands leaving £32 billion of profit growth behind (WPP Media, The Growth Gap)
Les Binet and Peter Field, The Long and the Short of It (IPA Databank, 996 campaigns)