Your next customer might start their product search in a conversation
“Which hiking boots are right for a multi-day trek if I have wide feet and want to spend no more than 180 euros?”
Anyone who asks like this already states their requirements, intended use and budget. AI assistants can turn this into selection criteria, compare products and prepare a purchase decision. Part of the advice therefore happens before anyone visits your online shop.
For retailers, this changes a key task: your products need to be clear and easy to find wherever the selection takes place. And you need data that shows which of these new paths actually lead to orders.
Online retail is growing – but not every shop is growing with it
According to the German e-commerce association bevh, online retail revenue in Germany rose by 4.3 percent in the first half of 2026.[1] This points to a market recovery, but says little about how an individual shop is performing. Product range, competition and the profitability of customer acquisition remain decisive.
Three changes shops should prepare for
Search terms are turning into concrete requirements
Customers still search for brands and items. On top of that, longer, situation-based queries are emerging. For your shop, this means: alongside product names, clear attributes, sizes, materials, compatibility and suitable uses all count.
Good product information should support both.
Part of the selection is moving to before the shop visit
Google continues to develop AI-powered shopping and features for agentic shopping journeys. Which features are available depends on the market and the respective rollout.[2]
This has an important consequence for analysis: fewer visits do not automatically mean less demand. If customers arrive with a narrower selection, visit numbers and purchase probability can develop differently. Whether this is happening in your shop is something your data has to show.
Product data is becoming part of your sales foundation
Prices, availability, variants and product attributes should be consistent across your shop and connected feeds. Google explicitly highlights how important the product data you provide is for AI-powered shopping experiences.[3]
Tracking has a different job from the product feed: the feed describes your offer. Tracking measures the observable response to it. Only together does this information help you tell visibility and commercial success apart.
What your tracking should answer in future
- Which identifiable sources bring visitors to your shop?
- Which products are viewed and bought?
- Where do customers abandon their purchase?
- How are order value and the share of new customers developing?
- How much of the reported revenue is left after cancellations and returns?
Keep in mind: not every AI recommendation leads to an identifiable click. A customer can get advice and later go straight to your shop. Website tracking alone cannot fully reconstruct that earlier contact.
That is why you need several perspectives:
These perspectives complement each other without having to deliver identical numbers.
Sources
- bevh: Erste Halbzeit im Onlinehandel: Kaufneigung der Deutschen kehrt zurück (in German), 6 July 2026. bevh.org
- Google: Universeller Einkaufswagen und weitere Möglichkeiten fürs Shopping (in German), 19 May 2026. blog.google
- Google: Driving retail sales in the age of AI, 8 April 2026. blog.google
Forecasts and assessments are labelled as such. Information on features and their availability reflects the state of the sources cited.