There is a line that has done the rounds in marketing circles for more than a decade, and nobody is quite sure who said it first: Google knows what you search for, social media knows who you know, and Amazon knows what you buy. It explains why Amazon built an advertising business worth more than most global media companies. 

Last week on Friday, at Glovo Kenya’s Future of Commerce Summit, two of their executives spent an hour setting out the Kenyan version of that line. Glovo knows what its customers buy, how often, in which neighbourhood, at what time of day and what they searched for before they bought it, and it now wants brands, banks and agencies to pay for access to that moment of decision. This post is about what it is selling, what the numbers behind the pitch do and do not prove, and what a Kenyan brand manager should ask before signing.

A US$200 Billion Business That Delivery Apps Have Discovered

However, some global context first. WARC’s Future of Commerce Media 2026 report puts global retail media at US$200.4 billion this year and US$223.4 billion in 2027, when it will account for 15.2 per cent of all advertising spent anywhere. 

It is the fastest-growing part of the advertising industry, and it is also maturing: WARC expects growth outside Amazon to slip into single digits next year, and it warns that Amazon, Walmart, Home Depot and Macy’s now serve more than 20 ads per page, with a backlash waiting if monetization degrades the shopping experience. Amazon holds 78% of American retail media and more than two-thirds in Britain, France, Germany, Italy and Spain. Retail media is a scale game, and the scale players win it.

Delivery platforms are the newest entrants and they are moving fast. Uber’s advertising business has passed US$2 billion in annualized revenue, growing more than 50 per cent a year across rides, Eats and in-car screens, and is piloting off-site ads that use Uber’s signals to reach people on Meta and Google and then tie the exposure back to an Eats order. 

DoorDash combined its ad products across DoorDash, Wolt and Deliveroo in June 2026 into a single platform with more than 400,000 advertisers and pushes its ads to search, social and the open web with closed-loop measurement. 

Delivery Hero, which owns 94% of Glovo and is itself the subject of Uber’s US$14.8 billion takeover offer, reported its European AdTech business growing 32% year on year in its second quarter. Advertising is the highest-margin line in any delivery company’s accounts, and every one of them is building it as fast as it can. That is the business Glovo has brought to Kenya.

The Global Product Set: 4 Goals & 4 Products

Marco La Porta, who leads partner revenue for Glovo globally, described the future of advertising on the platform in three words, intelligent, beyond visibility and full funnel, and then did something more useful than most ad sales pitches: he mapped a brand’s marketing goals to specific products. Branding and awareness, discovery and direct search, acquisition and conversion, loyalty and retention, with the brand in the middle and a product for each.

Smart Placements are Glovo’s automated native ads, sponsored store cards served across the customer’s journey on the home page, discovery walls and widgets, bought on cost-per-click with automatic bidding. The numbers on the slide were a 13.5x average return on ad spend year to date and 160 million served impressions. Keywords is search advertising, the restaurant or store appearing first for the search terms it has bid on, again on CPC auto-bidding with keyword-level reporting, with a claimed 15.4x average ROAS and a 20 per cent average click-through rate, which Marco described as the strongest CTR slot in the app.

Display Ads are the premium banners on the home page and the top of the food wall, bought on CPM with multiple creatives per campaign and sold to global brands; the slide carried Burger King, KFC, McDonald’s and Domino’s. Coming soon is Off App Programmatic: Glovo audiences, segmented by cuisine preference, order frequency and order value, reached on websites and social media in display, video and audio formats, with attribution connecting those off-app impressions to orders. That last product is the one to watch, because it is exactly what Uber and DoorDash are now doing, and it turns a delivery app from a place to advertise into an audience to buy.

The Number That Matters: 68% Of Searches Are For A Category. 

Ednah Gacheri, who leads brand advertising for Glovo in Kenya, started where every good retail media proposition starts, with the shopper’s moment of decision. Retail media, she said, is marketing to shoppers at the moment they buy, in the aisle or in the app, and that moment is where brands win or lose the sale. 

She asked the room to think about its last grocery order and whether it knew the brand it wanted before it opened the app. The answer, from Glovo’s own data, was the single most important number of the session. 68% of searches on Glovo Kenya are for a category, not a brand: milk, bread, soap. Whether they search or scroll, most Kenyan shoppers decide inside the app, which means the brand that is visible at that moment takes the sale from the brand that is not.

Her second point was against the media most Kenyan FMCG budgets still go to. “You know how many people saw your ad. Do you know how many bought?” Above-the-line media, television, radio and billboards, can estimate who might have seen it and cannot show who bought because of it. 

On Glovo the ad and the purchase happen in the same app, so, in her phrase, the loop closes from impression to sale. This is the entire proposition of retail media in one slide, and it is the reason Amazon’s advertising business exists. It is also, as I will come to, a claim that requires deeper interrogation than the slide gives it.

The Supermarket On Your Phone 

The smartest part of the Kenyan presentation was the translation. Every placement on Glovo was given a supermarket twin that a trade marketer in the room would recognize from a Naivas or Carrefour negotiation. The homepage banner is the front-of-store display. Category captaincy is the eye-level shelf. The category banner is the cross-category display, Royco in the vegetable aisle. Cart upsell is the impulse rack at the till. Samples in the bag are in-store sampling. Same spots as in store, Ednah said, except on Glovo every one tells you what it sold.

The products are then split by what you pay for. Awareness placements are sold on views: homepage banner, store wall banner, in-store banner, category banner and two tiers of order-tracking placement, the screen a customer stares at while waiting for the rider, which is best for launches, new flavours and brand recall and measured in reach, impressions and new customers. 

Sales placements are sold on taps: sponsored search, the first product the attendant hands you; category captaincy, owning the eye-level shelf; and cart upsell, the item suggested just before checkout. Those are measured in orders, sales value in shillings and return on every shilling spent. It is a tidy taxonomy, and it maps directly onto the budgets a brand already has for shopper marketing and trade promotion rather than asking for new money from the digital line.

Stop Broadcasting, Start Targeting

Audience targeting is now live, and this is where the data behind the line at the top of this post becomes a product. A brand can choose to reach shoppers who have never bought it, who shop its category, who are under 35; or shoppers who have stopped buying it or have been away for 30 days; or high-basket shoppers, Prime members, its loyal buyers, and, for a bank, its own cardholders. 

Glovo’s claim is a 2.3x return on ad spend for targeted campaigns against untargeted ones on the same budget. The creative follows the targeting: a billboard shows everyone the same ad, Glovo shows each shopper a different one, a party pack to the household hosting the match, an ice-cold energy drink to the 3pm slump, 20% off a first pack to the person who has never bought the brand.

Two products are new. Video ads put a brand’s existing television commercial inside the app, one tap from the basket, which for the FMCG companies that have already paid for the production is close to free inventory. 

Smart Boost is an AI tool that finds the words shoppers actually type for each product and places the sponsored listing against them automatically, which is Glovo’s answer to the keyword research a brand would otherwise have to do itself. The door is wider than it looks. 

Everyone can advertise on Glovo, Ednah said: brands sold on the app tap through to the basket, brands not sold on it tap through to their own website, banks and fintechs tap through to a card offer, and media agencies get one plan for every client. 

A campaign is set up once and runs across Glovo, the partner supermarkets and any store selling the product, with one report covering every sale from every store. The reach claim on that slide was 12 towns, 58 per cent of urban Kenyans and more than three million app sessions a month.

Cyka Fresh & The 9.6X Question 

The case study was Cyka Fresh, a produce brand that, by Glovo’s figures, earned KES 9.6 back for every shilling spent, took 45% of its August 2026 sales straight from ads, and grew sales 130% on August 2025. Every shilling traced from the ad to the sale, the slide said. The session closed with a summit offer of 20% bonus media for any brand, bank or ad agency booking its first campaign by the 30th of November 2026. 

However, every media performance number was Glovo’s: a year-to-date average, a platform-wide claim or a single case. None came with a base, a holdout group or a definition of what counts as an attributed sale. The question Amazon’s advertisers learned to ask the hard way is incrementality: of Cyka Fresh’s 45% of sales attributed to ads, how much would have happened anyway because the shopper searched for fresh produce and Cyka Fresh was already on the shelf? 

Closed-loop measurement proves that an ad was seen and a purchase followed. It does not, on its own, prove that the ad caused the purchase, and the difference between the two is the difference between a 9.6x return and a much smaller one. The advertising slide claimed reach to 58% of urban Kenyans which is almost certainly the share of the urban population living in the 12 towns Glovo serves. 

What To Buy First & What To Ask 

If I were a Kenyan brand or marketing manager with a trade marketing budget and an afternoon with Glovo’s sales team, I would do four things:

1. Buy sponsored search and category captaincy before anything else, because the 68% category-search figure means the decision is being made in the search box and on the category page, and those are the two placements that intercept it. Awareness banners can wait until the sales placements have a baseline. 
2. Bring the television ad, because video inside the app is the cheapest incremental use of a production you have already paid for, and bring it to the under-35 and never-bought segments, where the targeting claim has most to prove. 
3. Ask for the numbers behind the numbers before signing: the sample size and time period behind 13.5x, 15.4x and 2.3x, the attribution window and what counts as an attributed sale, and whether Glovo will run a holdout test on your first campaign so that the ROAS you report to your leadership team is incremental rather than observed. 
4. Finally, ask what the frequency caps are, because WARC’s 20-ads-per-page warning describes a destination that every retail media network drifts towards and that would damage the very moment of decision the whole offer is built on.

The larger point is where this is heading. Off-app programmatic, when it arrives, means Glovo’s audiences will be for sale on Meta and Google, not only inside the app, and the closed loop will run from a social feed to a delivered basket. 

Uber’s imminent takeover of Delivery Hero, expected to close in the second half of 2027, means that Glovo Kenya’s advertising business will in time be part of a group that already books more than US$2 billion a year in ads and is the most aggressive of the delivery platforms in selling its audiences off-platform. 

The retail media offer Ednah presented on Friday is the first draft of something that will be considerably larger, and considerably more sophisticated, within two years. The 68% is real and the moment of decision is real. 

The brands that learn now how to buy it, and how to measure it properly, will be the ones that benefit when the audience is three times the size. Google knows what you search for and Amazon knows what you buy. In Kenya, as of this past week, Glovo would like you to know that it knows both.