How we search and shop online has become more complex. People browse and compare, generating more touchpoints and data than ever before. The challenge is drawing out actionable insights from this ever growing dataset of modern customer behaviour.
The challenge is decidedly more complex for the retail sector. Retail marketers not only have to account for variables like time of day, device and messaging — they need to consider things like rotating product portfolios and categories, stock, distance from a buyer, and discounts. And this usually changes on a daily, if not hourly, basis.
In this context, machines can do the heavy lifting. But according to the Boston Consulting Group, only 2% of businesses are true “multi-moment” marketers1, meaning that they use factors like attribution, automation, and audience insights to deliver results.
You’ll find some multi-moment marketers at the performance unit at digital agency Mito. Mito has been using machine learning and automation to create effective campaigns for their clients — getting the bots to do the heavy lifting and putting data to work.
Be more relevant to the right customers
With 66 brick-and-mortar stores, Euronics is one of Hungary's leading electronics retailers. The retailer’s online store was launched over a decade ago and the team turned to Mito for new ways to increase Search revenue and improve ROI.
The challenge was not only reaching the right person but also showing them an accurate reflection of the website’s current inventory. With over 15,000 products that are updated on an hourly basis, it was near impossible to manually reflect product updates and account for variables in Search campaigns.
Euronics worked with Mito to launch a campaign that would reach relevant users with more relevant products and reflect inventory in real-time. Dynamic Search Ad campaigns were created and segmented by product category with custom labels in the page feed so products would be shown to relevant people. The top performing products were also segmented into a separate campaign with a higher budget so more traffic could be sent to best sellers.
Using page feeds that specified URLs, machine learning could determine which ads to show and where to direct people on the website. Ads are updated with new products and ads are automatically paused when a product is out of stock. This means that when a user clicks on an ad, the retailer can guarantee the product is in stock — avoiding disappointment and increasing the chances of conversion.
Custom scripts also automatically exclude keywords that are run by non-DSA campaigns and the team uses the DSA campaigns as a basis for keyword research.
The strategy paid off. Revenue from Search campaigns has increased by 336% and the number of online transactions has grown by 247%, while overall ROI is up by 300%. Machine learning is a powerful technology but it needs marketers to test and drive it. Get started today with these pointers: