We ran an Ozon ad campaign with a partner and closed the books: about $25,000 spent, $200,000 in sales, and overall ad cost share (ACOS) down from 9-10% to 6.7%. Below are the conclusions that changed how I think about advertising at large budgets.
Update. Ozon's ad cabinet no longer has a separate "push to top" format. What is available now is cost per click (auto strategy and manual strategy) and cost per order.
Budget scale changes the conclusions
Running ads on a small budget is one thing. Spending $25,000 over 40-45 days gives you completely different conclusions, different data and different insights. Some of them fully changed my understanding of advertising on Ozon.
This is not about ad settings - I have plenty of material on settings. This is about strategy and understanding. When you spend that much, you start to see advertising differently.
Ads amplify strong products
First conclusion: cutting the ad budget can sometimes produce the opposite result - higher sales and higher profit. Second: the most expensive mistake in advertising is usually not about the advertising itself. Third, and probably the main one: ads do not show how well you can set bids and tweak settings.
Ads show how strong your business, your products and your store are. If the numbers are weak, if there are questions about the store and the products themselves, the problem is usually inside them.
Advertising only sells strong products effectively. Ultimately it strongly amplifies what already works. Promote weak products and you get a fairly high ACOS. Promote strong products with good metrics - high conversion, high click-through rate, many reviews, strong ratings, quality infographics and quality SEO - and ads deliver a low ACOS. Promote weak products with poor conversion and unattractive design and ads deliver a high ACOS. That is the first conclusion from working with such large campaign volumes.
Do not scale weak positions
The second conclusion follows from the first: a common mistake I see in audits is scaling weak positions. Sellers come to me and say: "I run ads, my ACOS is high, I am not making money, it reaches 20-30%, what should I do?" We start the audit and usually find individual positions that deliver a low ACOS - but the sellers never pay attention to them.
They push other positions that deliver a high ACOS. Yet the most effective strategy is to strengthen what is strong and scale the low ACOS of that store.
After running budgets this large, I concluded that sometimes cutting ad spend is one of the most effective ways to increase sales. That follows from the first and second points.
The point is that we have top positions with a low ACOS and positions with a higher ACOS. Overall it is more effective to promote all products. But if we are not ready to put a lot of budget into high-ACOS positions, we can reduce promotion spend on them.
We do not push them at full throttle or set maximum bids. We find tools with minimum bids that give minimum spend. The products still run in ads, still get boosts for using advertising and still get promoted one way or another.
They clearly do not get the maximum amplification and maximum effect from ads. But they are still in ads and rank better in Ozon organic search.
How to find a reasonable bid
The fourth lesson: there is no single correct bid. I am often asked: "What bid is considered normal?" We have a target spend, ACOS, push to top, average cost per click strategy - each has different bids. And people ask me which one to set.
What I learned from spending a large budget: there is no such thing as a normal bid. There is no "this bid is normal and that one is not".
A normal bid is the one that delivers the maximum sales volume at minimum spend. Simple and logical: we spend the least budget while getting the most customers and the most sales.
The job of promotion on Ozon is exactly to find that ideal bid. On one hand, do not overpay. Setting an inflated bid is easy, and then you can tell yourself the ads are working, traffic is coming, sales are coming.
The hardest part is finding a bid below the recommended one. It saves money while delivering roughly the same level of traffic, maybe slightly less than top competitors who set maximum bids. That is the right thing to do when working with ads on Ozon.
Look beyond ACOS
The fifth conclusion: most entrepreneurs and sellers look only at ACOS. They watch whether their ACOS is this or that. But sellers often forget to look at margin, turnover speed, stock levels, LTV, repeat purchases and so on.
If ACOS rises to 25% and we panic and switch ads off, it will be far more expensive to get them going again. So advertising has to be managed not only by ACOS but by all the other key metrics: how much stock we have, how we turn inventory, what our margin is. It may turn out that we pay for ads and work at zero, or even at a loss.
Analyse dynamics, not single days
Lesson six comes with experience. The mistake I constantly see in consultations and audits: entrepreneurs analyse ads on daily statistics.
A dip today, a drop tomorrow, 12% one day, 20% the next, 6% the day after. And people start to panic: ads fell, what do we do, ads rose.
Look at the store we run with a partner. There are days with a good ACOS and days with a low one: 7%, 6%, 5%.
And sometimes ACOS rises to 14%. Often entrepreneurs, seeing high bids, high spend and high ACOS, slow ads down the same day: I will not run ads like that.
But you need to understand: periods of growth always alternate with periods of decline. Ads learn, algorithms learn, and in the moment we may get a higher bid.
Over the distance, however, it usually goes down. If you look at metrics in dynamics, you see the chart trending down and ACOS falling.
That only works because we treat advertising with discipline. We do not panic when the bid rises on some days and we do not make abrupt moves.
We give ads time to build up their work, and in return we get stable metrics and stable growth dynamics.
Ad formats and strategies
Lesson seven from this result: Ozon has several ad formats. Cost per click, search and recommendations, cost per order, cost per order for all products or for selected ones. There are quite a lot of formats.
Cost per order is straightforward. Let us look at cost per click. Within this format you can set different ACOS targets, meaning different bids.
There are push-to-top options across four strategy types: top 30, top 20, top 12, top 4. There is an auto strategy for search or for search and recommendations. There is average cost per click for search or for search and recommendations.
In practice there are a lot of strategies.
We ran many tests across different stores. All else being equal, two formats work best. First, cost per click on search and recommendations.
Second, cost per click on search. These are the manual bids we set and can influence. Yes, you can launch an auto strategy.
But we have always argued, and will most likely keep arguing, that manual bids work best. As long as our favourite Ozon does not switch them off.
Search vs search and recommendations
Which of the two formats works better? I am often asked: what is better, search or search and recommendations?
Search is a format where ads appear only in search results. A person types a query on Ozon for your product, and your product appears in search results in some of the first positions.
Search and recommendations is the same as search plus the recommendations page. Under every product there is a "Recommended for you" block. Your buyers browse competitors' product pages, study them, and somewhere in those pages they see the "Recommended for you" block that recommends your product.
Buyers click and order your product. These are two very different formats.
What does search and recommendations deliver? Impressions are many times higher - in this case roughly 20-27 times more than in the search format. But ACOS on search and recommendations is 6.6%, while on search we get ACOS of 2.2%.
What you need to understand about these tools: they deliver different metrics and different economics. The search format gives better economics, ACOS is always lower than on search and recommendations, but it also delivers far less traffic.
If the goal is a minimal ACOS, to stay within a small ACOS, we always launch the search format. If there is room to promote more aggressively and actively, we add search and recommendations - we reach more audience, get more impressions, more traffic and more customers. Yes, this will raise ACOS, but you need to track that it stays within your unit economics.
You can trade at zero or at a loss, but what is the point? You can push turnover to tens of millions, but what is the point if there is no net profit? So we always work carefully from the economics and margin we have set and do not exceed them.
What to do
Advertising on Ozon requires discipline and attention to economics. Strengthen strong products, test bids and formats, and watch the dynamics. That is when large budgets bring profit.
