Much is to be said about the new wave of buying and selling throughout the greater digital ecosystem. As e-commerce continues to transform the way businesses and retailers alike generate income, the industry must look to divide programmatic buying into two segments with different directions and futures, writes George Levin, co-founder and CEO, GetIntent, exclusively for RetailTechNews.
The first segment to focus on is in regards to brand awareness. As brands’ budgets continue to skyrocket, with respect to digital advertising, the importance of accurate targeting becomes essential for media buying. Specifically, data quality and accessibility are the growth drivers for this segment, with only the exception of massive FMCG brands – where pretty much everyone could be a target audience. For these brands, reach is the most important metric. They need to be 100% sure that their target audience saw the advertisement a precise number of times.
The second segment is e-commerce, where programmatic is much more performance-oriented, with very precise metrics, such as cost of purchase or registration. We can see another level of complexity when it comes to brick-and-mortar retail with online presence (which refers to pretty much all retailers nowadays) due to access to offline data. This is an area where programmatic buying becomes part of a more sophisticated martech infrastructure. As programmatic is driven by data, the more data you have – and the higher its accuracy – the more sophisticated your media buying could be. First-party data is the most accurate and most reliable, not only for targeting, but also for building look-a-like models. Thus, retailers have a serious advantage of getting the most out of programmatic. Let’s go through the typical programmatic buying strategies used by retailers, and necessary requirements to execute these strategies.
The most important part of ad tech is AI. Retargeting for e-commerce is probably the most comfortable environment for AI in programmatic. Based on historical behavioural data (such as browsing and search history), AI may not only predict which user has the highest probability of online purchase, but also predict what goods are most likely to be purchased. As a result, you will see banners, not only with products you browsed and didn’t buy, but also with products you most likely will end up buying (based on behaviour of similar users). Retargeting is still the easiest strategy.
The next level is using CRM data (both online and offline) to adjust bidding in real time. Each user gets a score (equivalent to the probability of buying something) and a recommendation engine picks the most relevant goods for each user. Basically, this is the same idea as retargeting, but with more sophisticated execution. A retailer’s DSP (internal or external) would make a pre-bid call to its CRM/DMP to get internal user scoring and use it to adjust a bid. For this strategy, there are three drivers: the amount of data per user; the accuracy of integration between the bidder and the CRM/DMP (latency is one of the issues); and the quality of the AI algorithms (how accurate they are in predicting probabilities).
Brick-and-mortar retailers may now collect more data on each consumer as they can access transactional data from the offline world. Two years ago, these conversations were mostly for big industry conferences, yet today it has become a reality. Intense competition throughout both the ad tech and martech ecosystems has pushed companies into very focused areas, where new competition has appeared almost immediately. This narrow focus, combined with the competition, has created very efficient products in every small faction of ad tech.
There are now plenty of data marketplaces that make it easy to buy and sell data. There are companies that align both the email addresses and mobile numbers with cookies, and companies that match devices to better understand the customer journey. For example, if you searched for something at work from your laptop, saw an ad on your tablet, and eventually completed the purchase on your mobile device, there are validation products that measure the quality of traffic and prevent fraud – and, finally, programmatic platforms that tie all of these component parts together and adjust bidding accordingly. The future of programmatic for retailers will be in harmony with the continuous merging of ad tech and martech. Retailers will also be looking to take programmatic in-house to reign in further control of their advertising campaigns. Something that Amazon did a long time ago.