⬆️ Winners in nomination Search Innovation
Google Partner

How Google Merchant Center feed optimization has reduced the transaction price by 3 times

PPC specialist
Reviewer: Galina Liman
Head of Context Department
5.00 1 rating
How Google Merchant Center feed optimization has reduced the transaction price by 3 times


Online store of jewelry and accessories on Prom.ua.


To increase the number of transactions to the pre-war level (it is possible to express it in units as 300+ transactions per month).

Period of work

Since the middle of April 2022 and continued cooperation. In the case study we used the information from Universal Analytics from 01.04.2022 (the first month of resumption of advertising campaigns) until 31.12.2022 (we are using the data for 2022, as from the beginning of 2023 we changed the strategy, and the results will be published later).

Starting conditions

  • Keyword search campaigns and dynamic search ads by product page feeds have been set up in the account;
  • Performance Max on Merchandise Feed and Trading Campaigns are not available because the customer does not have a Merchant Center, and the site does not meet the requirements to participate in the program;
  • The budget for clicks was halved compared to the previous periods.

An average value, for which the target for the PPC channel has been determined, the efficiency has been assessed and dynamics of achieving the goal is calculated based on the values for the first two months of 2022:

  • The number of transactions: 300
  • Transaction price: 285 UAH
  • Income: 241 200 UAH
  • Budget: 84 500 UAH

Timeline of the project development

On February 24, due to the outbreak of war, the advertising campaigns were suspended indefinitely. At the end of March, when Kyiv gradually returned to normal life, the client, like many other representatives of the business, decided to resume business activities to the extent it was possible.

We can conditionally divide almost a year of work into two periods:

The period of caution: during the first few months, we gradually and cautiously tested, set up new campaigns, and increased the budget where we saw a positive trend.

Full mode — starting from July, when we managed to reach the target, we breathed a sigh of relief and initiated more experiments as well as allocated enough funds for the simultaneous testing of several hypotheses.

The period of caution (01.04.2022-30.06.2022)

We are used to rely on the statistics collected during the working process, relying on which we can prepare fairly accurate forecasts and plan our work for at least several months ahead. This time such an approach did not work out. The information about past wartime behavior patterns was of little use, so we had to proceed slowly, tracking values in the shortest intervals possible, trying to identify at least some patterns in user behavior in order to build up hypotheses based on the data rather than intuition.

Average values during the first stage:

  • Number of transactions: 180
  • Transaction price: 178 UAH.
  • Revenue: 146,000 UAH.
  • Budget: 32000 UAH.

Optimization activities which had the biggest impact on the results

  • Registration of a Merchant Center account and feed creation using the Google Merchant feature for automatic feed generation.

In the beginning, we noted that the client did not have a Google Merchant Center account, and the main reason was the fact there was no software developer who could set up the generation of product feeds. Understanding to what extent the effectiveness of contextual advertising may increase, following the launch of product advertising, we offered to use the automatic feed as an experiment. If you have an account that does not yet have a feed with items, you can use the automatic generation of marketable feed.

Automated feeds are using the Website Scan feature, which scans structured data and maps to receive information about the products placed on the site.

  • Running Performance Max based on the automatically generated feed.
  • Adjusting the budget allocation over the entire period.

Normally, in order to make a decision on any changes to the budget, we need to collect a sufficient amount of statistical data (a minimum of 21 days of continuous operation). However, given the situation of complete uncertainty, we gave up the conventional approach. Moreover, the new campaign showed good results from the very start, and within 10 days it had almost the same performance in terms of the number of conversions and price per conversion as the branded campaign. Therefore, we started the budget reallocation process much earlier, and by the end of the first period, we achieved sustainable results, the improvement of which only required additional budget during the following period.

Comparing the results of Brand and Performance Max campaigns for the period from 01.04.2022 to 30.04.2022

The period of active growth (01.07.2022-31.12.2022)

By the end of July, we almost reached our target of 300 transactions, and after discussing the results with the client, we decided to increase the pace of development.

Average values during the second stage:

  • Number of transactions: 462
  • Transaction price: 83 UAH.
  • Revenue: 416,800 UAH.
  • Budget: 39,000 UAH.

The optimization activities that had the biggest impact on the results

Replacing the automatic feed with a classic one. Automatic feed is definitely convenient, yet the lack of optimization options somewhat limits the capabilities of contextual advertising professionals. The client provided us with the access to the admin panel on prom.ua and we managed to set up the required fields and create the custom feed with all the options for optimizing the attributes that can be used with the client’s products.

The result of these actions was the growth of transactions by 46% and an increase in the revenue by 86% compared to the previous month.

Comparing the results of the first month using the optimized feed and the preceding period with the automatic feed

We optimized the new feed immediately while creating it, added additional attributes to the mandatory ones to describe products as precisely as possible, and made changes to the important attributes in terms of future traffic relevance.

Among the most important attributes that have a significant impact on the effectiveness of Google Ads feed-based campaigns are the [title] and [description] attributes. This is due to the fact that, in the absence of keywords, the system relies on the information from these two attributes to decide whether or not to display a product to a particular user.

We’ve developed a scheme where [title] can be generated using existing product attributes, so you can change all product titles directly in the Merchant Center using feed conversion rules, if necessary.

Furthermore, we have also prepared a “cheat sheet” for the content manager of the client, to ensure that the requirements of Google are taken into account when adding new products:

  • Length: 150 characters
  • The most important information should fit in the first 70 characters
  • The name shall include: name of the product, brand or manufacturer, style, characteristics, material, attributes of use.
  • Check the frequency of keywords before creating an attribute formula Name
  • Frequent keywords should be placed at the beginning of the attribute Name

We created similar guidelines for the other attributes as well. In total, we added 8 regular attributes and 2 attributes with embedded values to the standard set.

A key role in the significant increase in the effectiveness of the advertising campaigns was played by the level of client engagement: they regularly initiated promotional offers with substantial discounts in order to provide users with an additional incentive to make a purchase.The effect of these actions was reflected in a 25-30% increase in conversion rates.

Regular update of the products range in online store. In this case, the effect of high-quality communication between the agency team and the client is evident. As soon as certain products or categories from New section demonstrated a potential growth of sales – we sent a proposal to the client to put these products into separate groups and actively promote these groups on higher budgets.

If the client managed to meet an increased demand, we sold the offer right away; if there were any problems with the availability, we offered similar products or delayed the start for a few days until the issue with the availability of the required number of products was resolved.

Inspired by the results shown by the optimized product feed, we decided to use it not only in Google Ads, but also to run product ads on Facebook. The fields and attributes are universal, thus practically no additional resources were required. During the three months of targeted product advertising, we managed to receive an additional 244 transactions at 128 UAH.


Close cooperation and involvement of the client in the processes resulted in:

  • 3x reduction in the transaction price
  • Increase in the number of transactions by 41%
  • Increase of revenue by 52%

Target (average values for the last 2 months before campaigns were suspended) Result (average values for the last 2 months of cooperation)
Number of transactions 300 516
Transaction price 285 UAH 82 UAH
Revenue 241 200 UAH. 455 256 UAH
Budget 84 500 UAH 42 000 UAH

Project manager: Oleksandra Nesina

PPC specialist: Oleksandr Shylin

Head of contextual advertising department: Halyna Lyman

Do something to stop this war! Russians are killing our children and civilians!