E-commerce advertising on Google Ads, Bing Ads, Amazon, and eBay

Fixing errors when updating the main feed

It was beneficial that the client had a development team, part of which was located in Ukraine. This allowed us to communicate directly and quickly resolve issues related to improving the feed. Numerous tasks were written specifically related to its optimization.

We increased the frequency of status updates and fixed bugs where scripts seemed to be working but were malfunctioning — that is, where they were incorrectly connected or incorrectly integrated with the ERP system. We analyzed the logs in detail when updating the feed, looked for sources of errors, and tried to completely eliminate them on at least one feed — the main one.

After about two to three months, we managed to fix most of the problems. At least, no new errors appeared in the logs when updating products. We checked that all items had correct names, images, descriptions, and statuses so that they would pass moderation on both advertising networks and marketplaces.

This is, of course, standard work in preparing products for sale, but the statuses were much more complicated. This required reliable, robust, and fast integration with other warehouse systems responsible for inventory accounting. This was the main focus of our work.

Adapting feed to the required formats and working with marketplaces

The second important part was choosing a tool (built-in or third-party) to adapt the main feed to different formats. We opted for a third-party solution because in-house development would have taken much longer and been more expensive. Purchasing a ready-made service seemed like the rational choice.

We used a third-party service where we uploaded the main feed and then selected ready-made templates for specific advertising accounts — Google, Facebook, Amazon, eBay, and others. All of this was perfectly suited to our tasks. Using this tool, we created correct versions of feeds for the necessary destinations, adapting the format for each platform.

In addition to global marketplaces, we also had plans to launch advertising on local platforms. If the first goal is to work correctly on global platforms such as Amazon and eBay, then the second is to expand to local marketplaces that operate only within individual countries.

For example, Denmark has several local platforms that are virtually unknown outside the country. However, they are popular within Denmark and generate good sales volumes. These platforms also have their own technical requirements and feed formats.

Here, the ability to customize helped us — the service allowed us not only to choose a built-in template, but also to independently adapt the format of the new feed to the specific requirements of local platforms, creating it based on the main basic feed.

With the help of the selected service, we managed to achieve significant results in a relatively short period of time. The main feed began to work without errors. Through a third-party platform, we adapted it to different advertising accounts, uploaded it to new links, and overall, the system began to work much more stably.

Orders for non-existent products disappeared, as did errors during checkout when, for example, a customer selected a color or size that was no longer available. As a result, negative reviews such as “wrong product,” “wrong size,” etc. stopped appearing. Gradually, the store's rating began to grow — both on eBay and Amazon.

This, in turn, is critically important because ratings directly affect product rankings: the higher the rating, the more often you appear in search results, the higher the probability of getting into the top results and, accordingly, increasing sales. Therefore, working to improve ratings became one of the key areas of focus in this project.

After stabilizing the feed, we began launching targeted advertising campaigns on Amazon at the client's request. For example, we ran campaigns for clearance sales or seasonal offers. In other words, if there were limited quantities of goods left in stock, we created separate ad groups for them.

Overall, it didn't make sense to advertise the entire range on Amazon — it was too expensive and ineffective. But targeted campaigns for specific product categories showed good results, and we continued to use them.

At the same time, we started working with local marketplaces in Denmark and Norway. We connected several platforms, adapted the feed for them, and made our first placements. We had to work with each platform individually because they don't have their own self-managed ad network like Google or Amazon.

Everything there is based on personal agreements — you need to contact managers, discuss placement conditions, banner positions, and participation in joint promotions. We negotiated partner integrations, promotional placements, and branded blocks. This kind of work takes more time, but it creates a “local presence” effect and strengthens the brand's position in a specific country.

An interesting part of the project was finding employees with rare language pairs. In particular, the client needed support specialists who spoke Scandinavian languages. At the time we were working on this project, ChatGPT and other modern translation tools had not yet reached their current level, and some did not even exist.

Therefore, the client asked us to help find people who could handle inquiries in Swedish, Danish, and Norwegian. The task was not easy, but we still managed to find several freelancers in Ukraine who worked part-time and responded to user requests via email.

Since the markets were just getting started, the client did not want to hire a full-time support team right away. So this was a temporary but effective solution — remote employees helped process the first requests and gain experience for subsequent scaling.

A separate direction was the launch of advertising in Bing Ads. Everything went much easier here because by that time we had already fully figured out feeds and analytics. We simply duplicated the campaigns from Google Ads, adapted them to the Bing format, and launched advertising in three countries — Denmark, Sweden, and Norway.

Bing had not been used in these countries before, so the new traffic source showed good results with minimal testing costs. Management was carried out according to the same principle as in Google, so there were no serious difficulties.

Key points

To sum up, the main focus of this case study is working with feeds. You can increase your budget and test ads and creatives as much as you like, but if the feed is out of date, none of it will work.

When you have thousands of products in your ads, it is critical that the system knows exactly what is actually in stock, what sizes and colors are available, and which products have been returned or sold out. Only with accurate and timely data updates can you expect stable conversion and positive customer reviews.

In this project, it took almost three months to fix all the errors and establish accurate feed synchronization. But the result was worth it — the seller's rating increased, the number of returns decreased, and the share of successful orders increased.

And where there are satisfied customers and high conversion rates, there is also a return on investment and the opportunity to scale advertising campaigns.