Navigation precision – German ecommerce marketplaces

By Pavel Zaslavsky, Jan 24,2021

In terms of navigation precision, German ecommerce marketplaces vary significantly. Some score with over 95% precision on their navigation systems while some go into 70% area and some are not even measurable.

In our previous blog post, we reviewed six major German ecommerce websites, and described their navigation infrastructure. We looked at how many filters and attributes each one of these stores offers to customers, to help them refine their results and find the products they are looking for. 


In this blog post, we will survey the navigation precision of selected German marketplaces. Creating a well structured infrastructure is only the first step. After that, product data tagging is a big and necessary step to ensure the efficiency of your website’s navigation tool. Precision measures how much of the tagging is actually correct, or how many, of all the tags you added, describe the products accurately. 


In order to measure the precision of the navigation tools, we went back to the same websites (,,,, and, and visited the same categories as before (Women’s T-Shirts, Women’s Boots and Smartwatches). This time, we randomly selected 1-2 values from different attributes in each category. We then reviewed the results to see how many of them precisely match the values we selected, and how many do not. In each category we reviewed up to 100 products out of the total results. 


In the Women’s T-Shirt category in Amazon, and selected the “Red” and “Esprit” values in the Color and Brand attributes, respectively. We got in return 90 results, out of which, 5 were incorrect, either not red or not by Esprit, and 85 were correctly tagged. This puts Amazon on a 94.4% precision in this category. 


Something went wrong in Amazon’s Women’s Boots category. After selecting the “Long” and “Buckle” values in the Leg Height and Closure attributes, we reviewed the results, and found that out of 100 results, 26 did not match the values we selected. The mismatches either did not contain a buckle, or were not high boots. In this category, the precision stands at 74%. Think about it - one out of four results does not match the values we selected! 


Amazon did better in the Smartwatches category. Out of 84 results shown after selecting the “Voice Control” value in the Features attribute, 15 did not offer voice control. Which means the precision level in this category stands at 82.14%.


In the Women's T-Shirts category, we used the “Striped” value, out of the Pattern attribute, to refine the results. After reviewing a sample of 100 results, we found 23 results that were tagged incorrectly and did not show a striped t-shirt. At 77%, the precision level in this category does not bode well for the rest.


Ebay’s precision improved in the Women's Boots category, going up to 85%. After refining the results using the “Kitten” value in the Heel Type attribute, and surveying 100 random results, we found 15 products that did not have kitten type heels. While better than the Women’s T-Shirts, there is still room for improvement. 


We checked back on the Smartwatches category, and refined the results by focusing on the Armband Color and Features attributes, and selecting the “Blue” and “GPS” values, respectively. After reviewing the results, we were impressed. Out of 100 results, only 3 were incorrectly tagged and did not match the values we selected (they were not blue), which makes it 97% precision. That’s a job well done on tagging here.   


In the Women’s T-Shirts category in, we decided to narrow the results to see only green t-shirts, so we selected the “Green” value from the Color attribute. We received 72 results matching our refinement, but after review, we found 7 products that are not green whatsoever, and so, incorrectly tagged. This gives a 90.28% precision level. 


In the Women’s Boots we looked into the Toe Shape attribute, and selected the “Pointed”. Out of 107 results we received, only 8 did not match the pointed toe value. Another good performance by Otto at 92.52%.


Indeed, Otto does not offer too many attributes to refine with. But when we refined the results of the Smartwatches category using the Storage Capacity attribute and the “4G” value, all of the 40 results matched the value. That is an impressive 100% precision level. 


In the Women’s T-Shirts category, we used the “Blue” and “Spring/Summer” values from the Color and Season attributes, respectively, to narrow down the results. Of the 100 results we reviewed, 8 did not match either one of the values, which gives it 92% precision. 


In the Women’s Boots category we refined the results using the value “Black” from the Color attribute, and the “Stiletto” value from the Heel Type Attribute. Out of the 99 results we got back, 12 were either not black, or with a wrong type of heel. This makes a 87.8% precision for Zalando.


The Smartwatches category has a 100% precision level. That is good. But if you remember for the last post, they only sell 16 smartwatches. In this case, anything less than 100% would be bad. Selecting any value in the Color attribute will give you back 1-2 results, which luckily for Zalando, match the color selected.


To narrow down results in the Women’s T-Shirts category, we choose to see only yellow t-shirts, but selecting the “Yellow” value in the Color attribute. We reviewed 100 of the results we got back, and out of those, only 5 were not yellow, which gives this category 95% precision.


In the Women's Boots category, we narrowed down the results to see only red boots. Again, we reviewed 100 results, out of which only 7 were incorrectly tagged, for 93% precision. 


After refining the results by Brand (Garmin) and Color (Black) in the Smartwatches category, the website produced 50 results. Out of these however, only 39 accurately matched the values we selected, while 11 others did not, which gives Real only a 78% precision.  


Yatego does not offer any navigation beyond the category level. As mentioned in the previous post, there is a Frequently Searched list, but selecting any one of these tags will take you to a different category altogether. In Yatego, unless you are willing to browse the whole catalogue, you have to type in exactly what you are looking for. 


In terms of precision level, we can see that there is quite a variety. Sometimes, there are significant differences in precision between different categories in the same website. It is worth noting, however, that the number of products you sell in your store has an effect on precision. Since precision is dependent on tagging, the more products you have to tag, the more prone you are to human error. So if you have only ten products in a category, the precision level should not be anything less than 100%. 


As always, we recommend investing in a well structured content infrastructure, and then put in the effort of tagging your products to achieve optimal precision.  This navigation precision survey of German marketplaces demonstrates that the situation is far from perfect and sellers have the responsibility to improve it for their own benefit

Request a demo



    PhonePlease, indicate the full number. Ex.: +9720501234567