STEP 1: GETTING STARTED
One of this year’s top trending topics in digital marketing has undoubtedly been personalization. Chances are you’ve been part of or have listened to conversions that included popular phrases such as “personalization,” “one-to-one fits,” “individualized fit,” and even “artificial intelligence personalization.”
This all sounds like exactly what we all want right? Personalized, yet highly automated experiences for our users that lead to more engagement and higher conversions.
But how do you hit the ground running and start incorporating these things into your current optimization strategy?
To start, we recommend keeping everything simple. Focus on one or two of your webpages (not the entire site) – maybe a high drop-off page or a low converting page – and start small, strategizing on how to increase one conversion goal on that page.
For example, we might decide based on the data, that we want to run a test on the homepage, and to start, we want to see if replacing the current hero image with personalized content increases our click through rate. After we run this initial test, we will then take the learnings and implement the winning solution. But this is just where the fun starts…
STEP 2: TESTING
Next, we will take a look at our newly implemented changes and ask ourselves if we can make another change that we’d like to test. Maybe there is another on-page element to test against our new control to see if the same conversion rate goal can be increased once again. Once we have chosen how to iterate with a new test, we will continue this testing process until we have found a winning solution that combines a variety of iterated and promoted tests.
Now we have a fully optimized winning solution! But why stop there? Why implement one winning solution only to let it live this way forever?
What if you could continue to optimize upon this winning solution on an ongoing basis, continuing to improve an already successful implementation and continuing to push the needle of your original conversion goal? This is where personalization comes in…
STEP 3: USING DATA TO OFFER PERSONALIZED EXPERIENCES
Just because we have found a winning solution to increase our click through rates from the homepage, we now need to understand and agree that one solution is not a one-size-fits-all solution.
Your visitors make up a variety of demoraphics and shouldn’t be bucketed into groups based merely on one or two known traits. What appeals to one person may not appeal to another, even among users who fall into the same demographical category. This is where we now need to shift our focus and personalized fit solution.
So how do we take our winning solution and find ways to make it appealing to each individual user? Unless we know each of our visitors personally, we likely do not know what is going to appeal to them each specifically, but we can make some pretty accurate guesses based on our data as we start to learn more about our who our users are and their individual preferences.
Again, we want to start small, but this time we want to learn what different creatives and options perform the best with different combinations of various contextual data for our users.
Our partners at Monetate, for example, offer a new solution for running individual fit experiences using their Artificial Intelligence Platform Engine. Within this platform the engine looks at multiple factors like time of day, location, device, sex, age, income, weather conditions, and much more.
These are “out-of-the-box” filters to which you can then add in your own custom categories to be more targeted to your unique business needs or user profiles. These data points are being considered both individually as well as together by the AI engine.
For example, if it is determined that our user has a family and prefers viewing content that caters to their desire to spend time with their family, we can offer them a creative that shows a picture of a family on the beach combined with messaging around taking a family vacation in the Caribbean.
If this user clicks on this creative, the machine learns and then begins to find more users like this one to serve this creative to as well. This increases and maximizes our odds of a user meeting the goal metric and clicking through.
STEP 4: GETTING SOPHISTICATED WITH A.I.
What is even better about the level of optimization and personalization AI offers is that the machine will continue to learn about our users over time, and will adapt its learnings in real time to real situations. So, for example, if one of our returning users suddenly stops engaging with these family images, the machine will begin to recognize the change in behavior and ultimately adapt for continued success.
Because no two families (or users) are alike, we shouldn’t treat them the same and assume all users who have a family will react the same to a single family option. This is why the machine looks at multiple data points – such as location, time of day, device, last search, income, etc.. – and not just “family” alone.
As cool as this is, this doesn’t mean our job is done when it comes to personalizing our user experience. Here is where the most important job comes into play. It is now our job to learn from the reporting that the machine is outputting and strategically update our creatives and messaging to best cater to and accommodate the users, now that we are able to understand their unique personas and behaviors better.
We begin to learn more about who they really are and what they truly prefer to see. Maybe we learn that the only users searching the Caribbean in our segment are family members? Not likely, but if it was the case, the machine would start to indicate this to us by showing these datasets in the top performing options so we can adjust our creatives and offer more family-friendly images. Neat huh?
STEP 5: CREATING A CULTURE OF TESTING & PERSONALIZATION
Ultimately, we take this information with us to the strategy table and stakeholders when helping with anything from merchandising sales to coming up with new testing ideas, because now we not only have analytical data but also our personalized learnings data to help make those informed decisions on how to best treat these users as individuals along every step of their journey.
Need help with any of the personalization or optimization techniques discussed here? Give us a shout at email@example.com!
- On 21st October 2019