It’s time to optimise for Google as a recommendation engine

Ask anyone what Google is and the most likely answer will be that it’s a search engine, which is pretty hard to argue with. The tech giant’s most important platform is called Google Search, where people type in queries and get search results in return.

Pretty conclusive stuff.

Except the nature of search is drastically changing, as Google further integrates machine learning and artificial intelligence into everything it does. We now get personalised search results based on our location, user history and content preferences. We get content recommendations popping up on our phones and alerts for the latest news, sports results and travel updates.

The further Google develops its technology, the less we find ourselves searching. Instead, Google is becoming more of a recommendation engine than a search platform – and marketers need to start optimising for this now.

What are recommendation engines?

Recommendation engines are systems that turn user data into recommendations. We see recommendation engines in action on platforms like Netflix and Amazon, where shows and products are recommended to us based on our viewing and purchase histories.

Amazon recommends additional products based on those you’ve previously viewed and bought. We also see recommendation engines in Google Now, which suggests content based on what you’ve engaged with in the past.

Google, Amazon and Netflix all use machine learning to draw up user patterns and “learn” what kind of content or products are most interesting to us. More compelling recommendations mean more time spent on Google, more shows watched on Netflix and more products bought on Amazon. This is a big deal for tech giants like Google, where more time spent on its platforms equals more ad revenue.

For marketers, it’s about optimising for a more personalised experience.

Creating a personalised experience for everyone

As soon as that confirmation email for next week’s flight lands in your Gmail inbox, Google starts providing you with flight info and weather reports for your destination.

Once you arrive, Google uses your location to recommend local attractions, restaurants and other points of interest. Each recommendation is one less reason for people to open up Google Search and actively look for information.

It’s not only travel Google wants to help out with, either. The search giant will warn you about upcoming sports events, suggest the quickest way to get home after work and even give you ideas for what to cook at the weekend.

More and more, Google is replacing the search process with a personalised set of recommendations for everything we do in life. So where does this leave us in terms of optimising for a search experience where users don’t actually need to search for anything?

Optimising for recommendation engines

As search engines continue to make more recommendations, we need to track which kind of user interests they’re targeting and what suggestions they make. This is important because you need to pinpoint which recommendations are valuable to your marketing objectives. You may not care what traffic advice Google has to offer, but you might be very interested in local restaurant recommendations for people in your area.

Instead of keyword research, you can think of this as recommendation research.

Near me searches are a thing, and now think of “best” as a term

Google doesn’t only interpret user search behaviour and serve suitable recommendations, it also trains our behaviour. Think of “near me” searches, which took off rapidly as Google suggested appending the phrase to the end of searches.

This is based on location awareness and it short-cuts the need for the user to add names of locations into a search. You don’t even need to add “near me” if the context of the search indicates that you might be looking for something local based on an immediate need.

Think beyond “near me”, though. Another area that Google has identified as a search trend is “best”. As users of Google, we have learned that it’s annoying searching for something, like a hotel, and having to trawl through loads of results to make a decision.

We now tend to type phrases like “best hotels in Stratford” or “best WordPress plugin for xyz”. The sites that are getting a lot of visibility nowadays are review sites, comparison sites or aggregators of lists of the best in any category.

Optimise for the way Google formats recommendations

Next, you’ll need to assess how Google and other platforms format their recommendations. Are they linking to Google Maps, AMP content, a website or something else? You want to make sure your content is on the right platform and in the right format to be discoverable in these crucial moments.

It’s also important to remember people will still turn to search at some stage in the consumer journey. Maybe they’re after travel recommendations outside of the usual selection or product reviews after seeing a specific suggestion. Sooner or later, users are going to take the search process into their own hands and you want to be there when this happens.

We also know more of these searches are being made through voice search – which brings another set of optimisation challenges. Targeting conversational queries, trying to feature in answer boxes and optimising for search results that might only include one listing are some of the tasks to think about. Once again, the hard part is identifying voice results that hold real value.

Is appearing in an answer box for “what is luwak coffee?” going to generate the leads you’re looking for?

Unless you’re in a very niche section of the coffee industry, probably not.

Identifying opportunities to generate valuable leads from recommendations will be one of the biggest challenges we face as searching takes the backseat on platforms like Google. Luckily, thanks to machine learning, we’re reaching a point of data maturity which enables us to spot patterns in user behaviour and platform performance in a way we’ve never been able to do before.

This will be integral to optimising for Google and other platforms as they turn machine learning and artificial intelligence into a more personalised experience.

Read how we’re using machine learning ourselves

At Vertical Leap, we’ve been incorporating machine learning and AI into our processes for a few years now. Below is just one example of how we helped one of our customers, Sarah Raven, increase transactions by 194%.

How we used machine learning to sell more flowers