The project came to us at the stage of a developed product that wasn't yet on store shelves and was generally sold only on a few marketplaces with low turnover.
The task was to launch product promotion to the target audience while meeting strict KPIs on the target cost per unit sold, and to squeeze the maximum sales out of this through the brand's own website.
To start the promotion, our team chose Meta Ads, because it let us reach the target audience we needed and build interest and awareness of the product. Its specifics lay in a certain novelty of the technology and the brand: it was quite unusual, and hardly anyone was searching for such a product on Google.
The website and sales tracking
Since our task was to start selling products through the brand's own channels (website and social media), the first thing we looked at was the website. When the project first came to us, the site was in pretty bad shape: built on WordPress with a free theme, with lots of errors, broken layout, poor design, freezes and a cart that didn't always work.
Conversion tracking was set up not on actual payment, but on clicking the buy button, which is incorrect and led to “fake” conversions being counted.
So first of all, we gave our recommendations for website fixes to resolve the key problems that could be fixed fairly quickly. We immediately reconfigured event tracking and connected the Conversions API for enhanced purchase tracking.
Sales through social media
One of the hypotheses to test was generating sales through the Instagram account. There are two common approaches in this funnel: campaigns that drive traffic to the account, and campaigns that start conversations.
During the first month of our collaboration, we tested both approaches while running website campaigns in parallel. Overall, the social media campaigns also brought sales, but by our calculations they didn't meet the required KPIs: a large share of people coming in through conversations had no particular intent to buy, and handling these conversations required a much more thoroughly developed lead-handling process with managers who could process them quickly and well. Meanwhile, the website campaigns running in parallel delivered much faster sales within the KPIs we needed. So we paused the tests and moved the entire budget to the website, to reach peak sales as quickly as possible.
How to showcase a new product
One of the challenges of the ad campaign was undoubtedly how to show the audience what our product actually is, since it was new to the market and it was too hard to show it fully in static images. The obvious solution was to create video creatives showing how the packaging opens, what's inside, what it looks like and how it's prepared — and to show all its benefits.
There are always two approaches to filming a product: expensive and cheap. Small startups obviously need the cheap option, so we brought a content creator onto the project — a single person without a big team, a studio or expensive equipment. Together with the client, we wrote a brief for the videos, sent the products for the shoot, and within a couple of days received ready-made video creatives.
And it worked very well: first, ad click-through rates went up; second, people started visiting the brand's social media, and we began getting organic sales from there. We believe the video was better at sparking people's interest, and our target audience began exploring the product more, including through social media.
New sales sources
In the first months of work, we found a balance between ad spend and keeping the required KPI on cost per portion sold. Further budget increases didn't allow us to maintain this KPI — we'll talk below about what to do about this and what we actually did, but first let's talk about TikTok Ads.
To diversify sales sources, try to reach a new audience and further increase sales, we proposed testing TikTok Ads. Since we had quite effective video creatives, this was an obvious move: we quickly prepared ad campaigns and launched them as a test.
Compared to Meta Ads, TikTok's cost per sale was about 20% worse, which was outside the KPI, so we temporarily paused this source — but only temporarily.
Scaling strategy
While working on the project, we ran into the KPI problem several times, when KPIs limited our ability to grow and increase ad spend. This happens quite often on various projects, and the explanation is that most business owners and marketers calculate the math strictly tied to the calendar month. We spent “this much” in March, earned “this much” in the same March, so “this” is the return. But this approach is wrong if customers come back again.
The point is that a project's math can depend heavily on customer LTV: whether customers come back for new purchases, how often they do so and with what order value. In FMCG, as in e-commerce, repeat purchases are quite common. They depend both on the specifics of the product and on how our marketing encourages customers to return, as well as on the quality of the product and service.
We suggested that the client check in their CRM how often customers return for repeat purchases within 3 months. It turned out that 20% do. Recalculating total profit and ad spend showed that the real return was much better than within a single calendar month. This makes it possible to raise KPI limits and invest larger budgets, acquiring new customers at a higher cost and earning more from them over a certain period of time.
After explaining this math to the client and proposing a switch to an aggressive scaling strategy, we resumed the TikTok Ads channel and scaled budgets. This allowed us to set new sales records every month and grow the project.
Over the course of our collaboration, thanks to all the solutions described above, we managed to scale the project's sales volume 5x. In our view, the key success factor was changing the way results are analyzed and calculated, focusing specifically on LTV and ROAS in the context of that LTV. Because with the old approach of monthly analysis, we would have stayed on small budgets and small sales volumes.
