How Machine Learning And Automation Give Brands A Boost In E-Commerce Advertising
CEO and Founder of CommerceIQ, the e-commerce channel optimization (ECO) platform that enables consumer brands to win in e-commerce.
As I write this, we’re all being inundated by news about Covid-19. It dominates our conversations because it’s affecting every aspect of our lives, from where we walk our dogs to how we should run our businesses. While I don’t have any strong opinions about dog walking, I do have my own perspective about running an e-commerce business now and in the months to come.
If you’ve read my other Forbes pieces (“The Keys To Success In An E-Commerce World” and “Is Grocery E-Commerce Here To Stay?”), then you know that I consider automation and machine learning essential to success in the world of e-commerce. The pandemic reinforces that opinion. Staying on top of demand spikes, inventory issues and advertising spend is challenging without automation during such volatile times. But these aren’t temporary concerns.
In fact, I think the virus has catapulted us at least five years into the future when it comes to e-commerce. Complexities that brands thought they could gradually learn to control are here now, all at once, and affecting every brand’s profitability. Advertising, in particular, requires a new approach. In my view, e-commerce advertising has changed permanently, and brands should use automation and machine learning to develop effective campaigns. Here’s a bit of history.
The Initial Approach
When brands took their initial steps into e-commerce (which meant selling on Amazon), they did what any of us would do: They continued using the same advertising agencies for print, television and outdoor advertising to handle their digital portfolios. Those agencies applied the techniques they already understood to the new world of e-commerce. What alternative was there?
The Interim Approach
As their understanding of e-commerce became more sophisticated, brands realized that traditional advertising techniques didn’t drive sales or promote growth. At the same time, the advertising world started to evolve, and agencies that specialized in e-commerce appeared. Those agencies understood the importance, for example, of keywords, and they took the first steps toward a better approach. But these interim steps are no longer enough.
Effective Advertising Now
To advertise effectively now, when we’ve been thrown into the future with no warning, requires new tools and a new vision. I see too many brands suffer because their approach is fragmented and misdirected. Different teams within a company are at cross purposes because they work from different data sets. Even more problematic is that marketers are far too focused on only using short-term metrics like return on ad spend (ROAS) as a measure for success when they also should be considering share of voice (SOV).
Algorithmic E-Commerce Advertising
We need a holistic view that optimizes the entire e-commerce channel.
With e-commerce, there are literally too many variables for humans to handle. This complexity is exacerbated by the speed with which situations can change. No longer coming up first in a keyword search or losing out to a competitor can happen in just a few minutes. An algorithmic approach to advertising can handle hundreds of constantly changing variables and give marketers and e-commerce executives the near-real-time information they need to plan their advertising campaigns strategically.
Don’t Just Rely On ROAS
I mentioned that brands shouldn’t just rely on ROAS as their measure of success. Taken alone, this metric is far too simplistic because it doesn’t take into account the many variables that actually impact profitability. As anecdotal evidence, one of my customers used an advertising service that relied on ROAS and didn’t take supply chain issues into account. They ended up promoting a product whose inventory was low. The ROAS looked great, but the campaign actually accelerated how fast stock was depleted. The item ended up going out of stock, orders couldn’t be fulfilled, and the product’s search ranking went into the basement. There was a short-term gain but long-term damage.
Do Rely On SOV
SOV is a far better gauge to determine if advertising spend is driving incremental revenue. This metric helps brands understand their market share against competitors at any point in time, from the consumer’s viewpoint. But it does more than that. It lets brands allocate ad spend to drive higher SEO rankings. So, how is SOV calculated? It’s determined by how well a product ranks for a set of relevant keywords. Can a human being do that, when even a single product can be associated with many, many keywords? Of course not. SOV measurement requires automation.
To sum up, achieving profitable, incremental growth requires:
• Focus on overall channel optimization, using SOV as a guide to measure performance.
• Automation and machine learning to get accurate, timely information and actionable recommendations.
• Cross-functional insights, allowing teams to succeed by providing the information they need to plan strategically for search and display across multiple e-commerce channels, like Amazon, Walmart and Instacart.
These bottom-line results speak louder than words.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?