From No Google Ads to KES 1.94 Million in Ten Weeks: A Launch Case Study
By Joseph Maina · · 4 min read
This business had never run a Google ad. Everything it knew about paid media came from Meta. Ten weeks after launch, Google Ads had recorded KES 1,942,243 in revenue from KES 125,422 in spend, a return of 15.5x across 339 tracked orders. Here is how it unfolded, including the first week, when it looked like a bad idea, and the campaign that did not earn its keep. The client stays unnamed. The numbers are real.
The starting point
The client is a Kenyan beauty retailer, already selling well through Meta. But Meta mostly reaches people who were not looking for anything at that moment. Nobody was catching the people already typing product names into Google, and those searches were going to whoever else showed up.
So the brief was straightforward: get in front of people who are already searching, track every order, and see if this can pay for itself.
Week one returned 1.4x. We kept going
One search campaign went live in the first week of August, on a small weekly budget. It returned 1.4x in week one. For a retailer, that is a loss once you count the cost of the goods, and many owners would have switched it off by then.
By week three, the same campaign was returning 27.3x. New campaigns need data before automated bidding and the account's targeting settle down, and judging one on its first few days is a common way to kill something before it works. It only made sense to wait because the orders were being tracked. We could see what was going on, and we were not just hoping.
Week one: 1.4x. Week three: 27.3x. Same campaign, same account.
Adding Shopping
Shopping ads began delivering in the last week of August. They show the product's photo and price in the results, so they suit a retailer whose products people recognise on sight. They went on to bring in 139 orders, at 14.9x.
In September, as returns improved, the monthly picture moved from 10.5x in August to 16.4x across September and into early October, and spend rose along with it.
What each campaign did
General product search was the star: KES 57,437 in spend, 176 orders and KES 1,032,685 in revenue, at 18.0x and KES 326 per order. These are people searching with clear intent, often for a specific product, and they are close to buying.
- General product search: 18.0x (176 orders, KES 326 per order)
- Shopping, all products: 14.9x (139 orders, KES 356 per order)
- Sister-brand search: 11.1x (11 orders, KES 745 per order)
- Local skin consultation search: 7.9x (13 orders, KES 789 per order)
What did not work as well
A local search campaign for skin consultations began in mid-September. It spent KES 10,257 for 13 orders and returned 7.9x, the weakest in the account, at KES 789 per order. That is more than twice the cost per order of general product search. It was worth trying, and the numbers told us plainly that it was not the priority.
A smaller search campaign for a related brand returned 11.1x on KES 8,200. It was respectable, but it had far less volume than the main campaign.
How far to trust these numbers
These are orders and revenue as recorded in the Google Ads reports, converted from US dollars at KES 129.89. Platform-reported revenue is not the same as money in the bank. Some of these customers may also have seen Meta ads, or might have bought anyway. A few weeks are missing from the weekly log, which is why the monthly chart adds up to slightly less than the account total. The ads were paused in early October.
Also, this is one retailer with established products and demand. A 15.5x return is not a forecast for your business.
What you can take from it
The lessons here apply well beyond beauty retail.
- Set up conversion tracking before launch. Everything here depended on it.
- Start with search intent, on a budget small enough to survive a slow first week.
- Do not judge a new campaign on its first few days.
- Scale what is improving, and compare campaigns on cost per order.
- Test new lines honestly, and accept what the results say.