Personalization Is a Team Effort, and Technology Is Only One Player

Many marketers treat personalization purely as a technological challenge. It's not completely off base to think that way. After all, nothing can be done without the technology that enables delivery of customized messages to designated targets. However, there is danger in that line of thinking. One may consider the mission accomplished when some fancy personalization engine is installed in the system. Successful integration of technology is a good start, but not the end. If you want proof of inadequate personalization, just look at your own inboxes and browser walls. Most so-called personalized messages are delivered by automated systems—and the majority of them reek of cold machines. Technology solutions are channel-centric, while consumers reside in multichannel environments. It is not like I would interact with a brand only through designated channels that are already set up with personalization engines; I would freely jump through many online and offline channels. In my journey to find a perfect TV set for my specific needs, I may search for reviews on the internet, click links on emails or texts, visit stores, compare prices on various websites, order online and pick it up from offline outlets. My journey may not coincide with technical flowcharts of channel marketers. Installing an engine in one of those channels is far from holistic personalization. The First Rule of Personalization: How to Not Annoy the Customer Data fuels the engine Personalization engines are only as good as data pumped into them. Data is literally the fuel, so it must be clean. Would you put unrefined oil into a fine-tuned sports car? Unfortunately, most consumer data is unorganized, uncategorized and disjointed. Reversing all that would be the primary goal of a customer data platform (CDP). The question is, do you have a CDP? Or are you just relying on channel data? Sure, personalization based on immediate behavioral data such as browsing history is useful, but it would hardly be the complete portrait of the target customer. That supports reactionary personalization, much like a salesperson finding out that a customer is looking for shoes and taking them to the footwear department. But what if you don’t have such information? Therein lies the dilemma, especially for new arrivals to your site. Where did they originate? What did they just browse? Which products did they examine at length? The thing about explicit data is it's great when you have it, but it can be very narrow. Sure, your personalization engine is ready to show different content to different individuals, but what if you have no idea what they want? That is when you must escape the myopic world of immediate and explicit information, and venture into all types of old and new data, possibly captured in your CDP or even external sources. To use such vast arrays of data effectively, marketers must employ advanced analytics such as segmentation and modeling. Customer segmentation done right As a starter, you may treat people in distinct segments differently. Surely, you would want to treat one-time visitors differently (welcome series or various upselling offers) from loyal high-value customers (special invitations and exclusive offers). Using various affinity models would further deepen the quality of engagements. For cross-selling or upselling, what should be the featured product for each target? If a person scores high in multiple affinity categories, which one should be used? For instance, I may score high for both travel-related products and cutting-edge electronics. How would you break the tie? Simply, go with the category with the higher score for the target. Comparing the merits of segments and personas deserves a longer discussion, but let me just summarize it this way: They are not mutually exclusive options. You must use them all, as you won’t be able to control the data flow at every stage of customer engagement. The final puzzle piece Now that I briefly covered display/delivery technologies, various data types and analytics options, the final piece of the puzzle in the holistic personalization game is the content. There are too many cases where personalization efforts fall short due to lack of content. What good are all those engines and algorithms if everyone gets to see the same offer and picture in the end? This, too, deserves a longer explanation just to cover effective ways to marry personas and content libraries. But the main message here is that this personalization game is definitely a team sport, and the content designers and providers must be onboard from the planning stage, not just technology and mathematical geeks who deal with toolsets and algorithms. I have seen too many instances where designers would say something like, “You want different versions of creatives for all these personas?” at the 11th hour before campaign launch. Don’t be that kind of company.