“Google Shopping now accounts for 65% for all Google Ads clicks and for 89% of non-branded Google search ad clicks of retailers”2019 report by Merkle
Managing, checking and curating thousands of images manually is almost impossible.
Google Shopping Ads can offer a substantial return on investment. They give your customers an opportunity to see your product and its price as soon as they start searching.
However, setting up Google Shopping Ads requires considerable time and attention. Your product feed needs to be 100% accurate to get the best click-through rate. And that’s simply not possible. Your team will make mistakes when spreadsheets contain thousands of data points.
This is where Artios comes in. We use deep learning to quickly and accurately spot problems.
Deep learned image processing doesn’t make mistakes
Google Shopping requires you to specify numerous attributes for each product. For apparel items in the UK and US, Google requires attributes on colour, age-group, gender and even size.
Inputing all these correctly manually is almost impossible. So what do you do? Check them all manually as well? Mistakes still creep in and you waste months of your marketing team’s time.
Instead, try Artios. We’ll use deep learning to quickly and accurately identify mistakes.
Deep learning can be trained to solve specific problems
Do you want to know if your Google Shopping Ads conform to your style guide?
We can train our image processing tech to correctly identify images that are lifestyle, flat, studio model, close up or wide. You can see which adverts no longer fit your brand and need to be updated.
Our deep learning programme can also spot attributes like: image style, garment type, colour, pattern, model gender, and many more.
We worked with a watch retailer whose watches were often incorrectly identified by brand and type. These watches were high value items and classification mistakes – caused by human error – were costing the retailer a substantial amount of money in lost clicks.
We used deep learning to extract brand information from product images and check them against the product database. Mistakes were swiftly identified and could then be corrected.
Artios saved the retailer at least 3 months work manually checking the product feed for each watch. This would have been a boring job that required considerable attention to detail. It couldn’t effectively be outsourced.