Beyond SEO: REO
Beyond SEO: REO
Recommendations Engine Optimization
No one buys a new product without a recommendation. 95% – just to avoid saying 100% – of us will consult ratings & reviews. These are central to any purchase decision.
A famous McKinsey paper states that 35% of Amazon sales are done through their world famous product recommendations engine, the ‘Customers also bought ..”. Not to mention Netflix movie suggestions and the always alluring YouTube right hand side ‘you may also like’.
Recommendations are everywhere and I would argue they are fast becoming the most important eCommerce tool. I would also agree that in a not so distant future, we will have recommendation engines so powerful that they will handily beat any human expert advice. They will know you better than anyone and will be able to predict your purchase behavior better than yourself.
Imagine for a moment that you are able to gather feedback at a global scale. Everything and anything which has ever been said about a product. The good, the bad, the ugly. Now imagine you can associate these attributes not only to the product but also to the reviewer. Someone living in New York will have a different opinion on the very same product than someone from central Utah or Uzbekistan for that matter: The reviewer matters as much as the product itself. That is particularly true for service businesses like restaurants. Your taste, price expectations and appreciation of a specific decor are largely influenced by where you live, who you know, and of course your own self being. That’s the reason why the Tripadvisors of the world now account for where you live, your gender, your income level and so forth to refine their suggestions.
What works well for services like restaurants also works very well for products. That’s why any ecom try to learn as much as possible about their customers as they can. And soon, we will see a much larger use of the next level of recommendations engines which until now had only been available to the limited set of data rich, R&D rich GAFA’s type of the world who have made their mission to get to know you better than you can possibly remember. As technology becomes more prevalent, the data (think digital traces turned customer profiles) will truly make the difference between who wins and who loses.
Better than your friends
A sophisticated enough recommendation engine is more accurate than any direct friends or expert recommendation. A sophisticated recommendation engine accounts for a very large number of customer attributes and accurately derives the most appropriate product. With thousands of products and attributes this is a task way beyond any human being ability. But with the correct models and the right dataset this is very achievable. The heavy use of neural networks (and all their subsequent derivatives) are transforming what used to be glorified filters into truly sophisticated, accurate, trustworthy systems.
I would even argue that you can push that reasoning all the way through medical studies and improve on the more classic approach. When you test a drug, you have a lot of parameters to account for and each of them can interact specifically with each individual. As we are going through Covid vaccination on a global scale, imagine a system that would know everything about each individual receiving the shot and gather any reaction to that shot from all these individuals. Assuming every patient shares their results through the process you end up with a massive dataset from which you can predict how the next person will react to the vaccine. You have enough data to predict extremely complex biological responses without having to fully know how the biology itself works. That is how sophisticated recommendations engines will work and why they will quickly become a central tenet of any commercial transaction.
From digital traces to trusted advice
From customers ratings and reviews on products, from blog, vlog and any other digital trace, recommendation engines are swallowing that massive information to accurately predict what product will work best for which customer with a very high level of accuracy.
While a search engine brings customers to an eCom site or a store, the recommendation engine is that last push that makes the conversion happen. Today, when we shop, we are still reading reviews – mostly contradictory reviews – and trying to make sense of them. We are getting a ‘feeling’. If we see enough positives we will go for it. If we read horrible things we will run from it. The next generation of recommendation engines will know enough about you to select the reviewers that match you the best. It will be much more accurate than this random reading where we can only scan a tiny piece of the relevant material. It will rely on powerful models to tell you with a high degree of accuracy if that product will work for you or if instead you should get that one.
That is where REO, recommendation engine optimization, becomes critical. SEO brings customers to the store, REO is the last push to conversion and more importantly, the trusted push to satisfaction.
In a digital world, once a user lands on a site, digital traces, queen among them reviews, are more central than anything else to convert. What is true online is also true in a physical store. Accurate personalized recommendations are the sales associate best weapon of surgical selling.
The same way SEO helps website discovery, Recommendation Engine Optimization (REO) helps product recommendations in a way we haven’t yet fully experienced. I am convinced that we will use customers gladly sharing personal information to further refine their profile in order to receive the best possible advice. When you visit a doctor you share all the info at your disposal to ensure the ultimate diagnosis. The most refine REO systems will be there to give you advice that resonates so well you won’t believe you didn’t think about it yourself.