Building Great Personalization: Components (Part 1)
Such as behind every great person there is a great woman, behind every great personalized experience there is a great marketing technology stack. In this series of articles, I’ll walk you through the components that you need to bring your personalization program to life and would describe their different types.
Component number 1 - Great Segment Builder
IMHO, a great personalization tool starts with a great segment builder which is the core of the technology. Because there are different approaches around what personalization is and is not (and I’ll make an article about it soon) there are different tools to meet those ideas but all of them are around the different ways to organize segments.
I’ve analyzed many types of personalization tools available on the market today: from the free one like Google Optimize to the most expensive with no free trial access like Adobe Target, from the most “user friendly” to the “write your own code” kind of guy. A huge disappointment and money waste could bring some “mid-market” kind of tools which promise you “the great results” but are not able to deliver to the buyer’s expectations. I put together my thoughts and divided personalization tools into 3 groups which would help you to clear up some of those expectations.
Level 1 tools
There are a bunch of tools (~90% of the market segment) that allow you to build segments based on users' web behavior only. They are limited only to that capability. Most of the time that is those guys who bring disappointment and wrong perception about personalization. They are not bad, though, and could play a vital role in the first attempts to run your personalization program.
They could help you to bring to life projects like:
- Campaign message synchronization — content adjustments based on the traffic source. For example, if you run your Facebook ads for pet owners and want to create a separate landing page experience for cat owners ads group and dog owners ads group, so you would want to embed the campaign’s audience in the UTM parameter and set up a rule to adjust the offer experience on the landing page accordingly to the group. Frankly, you can do it without an additional tool by just creating 2 separate landing pages and feeding your traffic to the different pages.
- Adjust landing page experience for different devices/browsers/screens. Quite helpful, but to be honest, who’s doing that in a manual fashion?
- Set up a rule to adjust experience based on geographic location. Helpful if your users are not using VPNs though ;)
- Set up a rule to adjust experience for previous web behavior, e.g. if a client abandoned a cart, retarget them on the homepage with the product they left in their shopping cart.
- Set up a rule to adjust experience based on google analytics’s affinity category, gender, organic/paid keyword they came through.
How does it work?
These tools are able to set up an internal user id, attach it to a device id, store it in a cookie and retrieve it at the right moment. Alongside the internal user id, it is able to read IP signals like location, provider, and data center, read device’s characteristics like screen dimension and model, read referral signals such as referring domain/URL, read UTM parameters and any other second/third-party information which comes with the tool like affiliation group, provided by Google Analytics.
They would be great for companies:
- with low customer retention (most of the e-commerce stores)
- with teams working on the awareness stage (growth hackers/brand awareness/demand generation)
- Start-ups who just started to win a market, actively growing their user base and trying to find a “right” message for the market.
Again, that’s all cool and works perfectly if you are just starting your personalization program. But buying those types of tools it would be better to adjust your expectations that it’s basically what they are capable of.
Level 2 tools.
There are not a lot of tools (7% of the market) that allow you to bring your own data into the system such as Zero-party, first-party and second-party data. (I promise to write an article about it)
Why is it important?
These kinds of tools would allow you to bring your personalization program to the second level to build customer segments based on your own data and web behavior data. Which is some level of personalization but not 1:1 of course.
You would be able to perform this type of magic as:
- Set up a rule to serve a specific experience for your existing customer whom you recognize and able to classify, for example, “high-value customer”, “silver-card holder”, “churning customer” etc.
- Mix your customer data with their web behavior data, e.g. “serve experience X to the high-value customers who are actively engaged with Z page.”
- If you really want to, you can create 1:1 level segments and set up separate experiences for each of your customers. (Hell no, I’m personally not doing that crap, but it’s possible)
- And, of course, everything mentioned on the previous level.
How does it work?
At this level, the tool should have the capability to:
- Assign an internal user id to match with a device id (e.g. UUID or “device-id”), store it in the cookie. When a user leaves personal contact details (e.g. email/phone number or “cross-device id”) on your website or app, it should be able to retrieve the internal user id from cookies and match it with the personal id (email). “Cross-device id” would be the unique user id which would work as a key-connector b/n your Customer database (CRM) and digital activity database (could digital channels like app, website, kiosk, POS)
2. Have a built-in id classifier that would allow you to build, store and retrieve the segments. Meaning, based on contacts classification available in your CRM (date of birth, last purchase date and last purchased product, marketing segment, sales segment, etc ), you would be able to upload this classification to the tool, and the tool would be able to retrieve this specific group of “cross-device ids” (global_id on the image) with the existing in on the server group of “device ids” (user_id on the image) to deliver the experience.
This level of tools would be great for:
- Demand gen teams who know how to acquire client contacts, but struggle with activation, low engagement, or bottlenecks.
- Customer resurrection teams who are trying to solve the churning problem.
- Companies with stable and long-term relationships with their customers and looking for a new way to increase revenue via cross-selling and up-selling.
Level 3 tools
Just approximately 3% of personalization tools on the market are capable of delivering the real 1:1 personalization experience. The core of the personalization technology here is in the AI/Machine learning components or ability to digest custom data models.
Why it’s important.
These tools don’t work based on a simple “if-then” rule as on the previous 2 levels but the core of the component is in a prediction such as “there is 95% chance that this guy would like this experience better than the rest of possible experiences.” It almost eliminates a human “guess factor” on the way to bring the right content to the right person at the right time through the right channel without compromise. Pre-set manual segmentation is surrendered to a real-time profile building. It acknowledges ever-changing human nature and works with it trying to organize the chaos.
How does it work
It analyses a massive amount of web data across thousands of users on the website, classifies them into groups with similar behavior, and learns about the engaging triggers. When a new user comes to the website, the algorithm would run a prediction (prior knowledge) that “this guy belongs to that group”, test the hypothesis, and spits a conclusion based on the next observed action (posterior knowledge). Every piece of data would prove or disprove a hypothesis about “this guy belongs to that group”. And all of this happens in a fraction of a second while this guy is browsing your website or app.
Great for companies
Needless to say, this type of tool would cost a lot of money. A lot. Or would require a lot of engineering, data science, and management resources if decided to build “in-house to cut ops costs”. But money is not everything ( sorry, mister Putin), companies should also have :
- stakeholders’ buy-in (I promise to write an article on how to bring innovation in your “old-bones” companies) on personalization is “a thing” and ready to “chip-in” resources, especially, creative resources and
- great technical and human foundation (promise an article as well) before those expensive companies would start charging you.
Level 3 tools are the tools of the true 1:1 personalization and the tools of how the future of marketing would look like.
There are 2 more great components necessary for great personalization, but I’m going to talk about them in my next article.
If you want to learn more about personalization and/or have a great project in mind you need a consultation on and/or just want to host a chat on the clubhouse platform, bing me in! https://www.linkedin.com/in/marinazub/
