Google “personalization.” Maybe you just did.
Did you add “digital” to that search? Because otherwise your results may be engraved crystal beer steins. Well, actually, hold on again… take a closer look at those glasses. Is that your initial on them? Yep, your search results for personalization have been personalized.
Keep scrolling and you’ll probably see “the future,” “critically important,” or our favorite, “your ticket to.” All noisy marketing-speak found in bold type on every vendors event signage or services page- and one that people don’t know quite what to do with. Like those “personalized” beer steins, it’s being etched onto everything and sold in a million ways.
If you’re excited but confused, it’s for a good reason. “A/B testing,” “Personalization” and “Optimization” have quickly become both unavoidable buzzwords and ubiquitous digital budget line items. For some, these are rapidly emerging practices. And for a very select few, they are foundational to everything they do. But for most, the term “personalization” persists as the most uttered and least understood term in the ethers.
If you ask 100 people what it is, you’ll get 101 answers. Today we’ll try and unpack what you’re hearing and walk through the basic “how do we’s” of personalization.
What we hear most is “we need to be personalizing,” then crickets. Crickets about something you believe is “critically important” is never good, especially when money’s involved. The following isn’t more jargon to drop, its our attempt to help you turn crickets into conversation.
The umbrella for all of the following activities. Optimization involves the data-driven improvement of a user experience and business performance of your website. If you missed it, our recent blog sheds further light on Optimization as an approach as it encompasses many activities versus being a singular service.
Any optimization effort relies upon the testing and validation of questions you’re asking about your product or site attributes, these are known as hypotheses.
The lifeblood and fuel of any and all digital services or improvements. All optimization activities begin with the claims you make about how to improve your business or customer experience. This is not cute scientific jargon, but rather how you’ll identify the foundational questions and assumptions driving your marketing plan, design comp, product features, merchandising efforts, copy changes and promotions.
Your hypotheses will help define if and how you construct personalized tests and experiences for your users. Think of them like blueprints. Every personalization effort begins with a hypotheses about how the targeted content will benefit the business segment or individual.
Why your specifically defined users will feel valued and rewarded. This term refers to the dynamic insertion, customization, or suggestion of relevant user content based on their implicit behaviour and preferences. Monetate’s “The Business Case Infographic for Personalization” provides some of our favorite data on the cost of ignoring customer relevance.
Personalization is significantly defined by the (1) size and characteristics of the segment and (2) the method for delivering the content to the segment. Personalization testing and segmentation are not mutually exclusive concepts – they work hand in glove to achieve your customer’s perfect fit.
Now, about testing… (now’s a good time to top off that coffee.)
Your high confidence, two-way mirror for viewing customer’s preferences. A/B testing, or split testing, is comparing multiple versions of a web page or pages (A, B and so on), in a controlled environment, to determine which produces positive results before you invest further. By measuring the impact that changes have on your metrics such as purchases, sign-ups, clicks, downloads, etc. against your control, you can ensure that your changes are likely to succeed and avoid or re-think those pursuits that do not bare fruit . Optimizely offers a great overview of A/B testing, and how it helps you move from “we think” to “we know.”
Multivariate (MVT) Testing
Like having several related petri dishes under a microscope. MVT is similar to A/B testing, but compares how a high number of variables and permutations interact with one another and against other permutations. (See Wired: “The Technology That’s Changing The Rules of Business”)
Fitting Things Together
Digital optimization is inclusive of both AB testing and MVT testing. Personalization is a way to target experiences that may or may not be tested. The formulation of your hypotheses will drive all of the above.
Personalization could thus be characterized by the method of testing segmentation or targeted attributes in conjunction with the method of targeting content or experiences based on the rules or definitions generated by the segment input or behavior.
So now that we understand the terms, how exactly do you move from hypotheses into effective segments? This is where even more debate lies.
Turning Questions Into Segments
Your team’s own hypotheses will open up a debate on what segment to target, how big the segment is or whether a segment is worth testing at all. Consider this – is it better to target a very refined segment up front or begin with a broader segment then use collected data to refine later? Should you manually create and test an experience against a segment or employ algorithmic methods to serve experiences based on machine learning?
How much, when, and how to personalize varies for each business. Ultimately the answer will depend on (a) how big your audience is and how big the definable segments are (b) how much you have already tested and optimized against the broadest segments (c) what resources you have to create custom experiences for each segment and (d) what team or software is guiding you through the process.
Really? That’s it?
Yep. For now. The purpose of this post was to unpack “personalization” – take some of the mystique out of it and to confirm the following:
- Personalization is a practice distinct from A/B testing, but it can also be a layer or logic that enhances an A/B test.
- Similarly, A/B testing a personalization effort can confirm whether the effort was successful.
- In all cases, the action should be driven by a clearly stated and prioritized hypothesis.
- The success of personalization efforts depends largely on (a) the size of the segment, (b) the ability for that segment to convert and (c) the cost and effort to target said segment.
While there are many ways to personalize and no singular “best way”, there is a way that is most pragmatic for your business. Stay tuned for our co-founder Ryan Garner’s upcoming post on our what we’re seeing work best and what the future looks like for the industry’s favorite buzzword.
If you’re spending a lot of energy or dollars wrestling with how to make personalization, testing and optimization accessible and practical for your team or your company… you’re not alone. Give us a tap and let’s tag team it.