How AI Enhances Personalized Advertising Experiences for Retailers
By Serge Ioffe
In the new reality of customer-obsessed marketing, all industries have a number of standard advertising creative guidelines to drive customer acquisition, loyalty and retention. When implemented correctly, meeting these guidelines drastically increases ad performance over standard, outdated generic ad copy, targeting and visuals. The result is a major improvement in the buyer journey – through more relevant content in advertising.
For retailers, effective AI-powered digital advertising mix should follow a model informed by at least four essential moving parts.
Most retailers are increasingly adopting a “modern creative” approach that showcases SKU driven content. The SKUs come from a catalog data feed, which allows a lot of interesting customizations, including sale prices, categorization clustering, collaborative filtering and much more.
A recommendation is based on many different possible variables, including:
- Retargeting (did you shop for those pair of shoes)
- Computed Interest (do you have interest in shoes?)
- General Merchandising Priorities (we don’t know if you want shoes, but here are shoes we want to sell now)
- Recency (we’ve shown you shoes for a while and you haven’t responded, so here are some socks)
- Profile (we saw that you shopped for multiple family members, so we’re going to add shoes from outside of your category)
Promotions overlay on top of SKUs or SKU specific sale information. For example: a running shoe. It used to be $100, which is now shown as a strike through price, and now it is $80, shown with a price that highlights the sale.
However, at the same time, we can compute that there is a promotion running for all shoes (for example “20% to 40% off until the end of the month”). If we have decided that a user is interested in shoes, we can look up the applicability of a specific promotion from a promo level data feed. Finding the promotion, the creative would then overlay its text and imagery on top of the rest of the ad. Importantly, this is subject to ongoing testing to determine the best layout or creative variation that best presents the information.
Retail Support/Store-driven Support
Retailers often have a physical presence, as well as a digital one. And if so, they want their ads to convey a call to action or information that will get the customer to visit the store. This requires advanced logic, as well as a data feed with store locations, knowing where the user is, and which store is closest. This system also includes a default creative if the user’s location is not within a set number of miles. In addition, the feed will have to take in any other targeting variables such as in-stock availability, retail price, promotional price, and maybe even specific scheduling to support the store during business hours.
Event Specific Badging
Most businesses have a sales schedule. These sales or events can occur in quick succession and need to be rotated out quickly. Standardized ads can have various “skins” or “badges” that are applied to them that can be used to quickly and easily apply sales event messaging on top of all other current messaging.
In conclusion, through AI and strategic use of data, retailers can create ads that result in a new breed of optimized advertising that serves both the seller and buyer. For example, by using the data sources listed in this article, the resulting ad will show the consumer desirable products that are available in their local store at a discounted price, that have a promotion going for an extra 20% off – along with badging and themes of a specific sale event running at the same time.