Road to Personalization
The marketing community loves buzzwords. One may say that some words just go viral. In the past, CRM was one. Server-migration (from mainframes) was another. Cloud computing – even among non-IT groups – has some magic power. Big Data has indeed been a big one the past few years, though it surely is losing its coolness, especially among data professionals. But, in some countries and communities, it is still gaining momentum. The latest one, I think, is “Personalization.”
Do you know how I get to find out how some words are becoming popular? The fastest way is to attend a conference and check out which session keywords are filling up the rooms. Attendance, like in the movie industry, is a sure way to measure the power of the keyword. We often see that some speeches and articles are not even remotely related to the word in question, but that doesn’t seem matter much. Everyone and their cousins start selling the word like it’s a magic potion that cures all. If you happen to come across a password to a goldmine, won’t you try it, too?
Once the word starts go viral, the power of magic starts to influence the real-life decision-making processes. Yes, I’ve been using every chance to debunk the mystery around Big Data, through this series and other opportunities. But I have to admit that those two words originally strung together for marketing purposes by software companies opened so many new doors to meetings and speaking engagements to which data geeks never dreamed of having access a mere four to five years ago. If you ask me what the best outcome of the Big Data movement is, my answer is that decision-makers, in general, became aware of the importance of analytics based on collected data. Analysts no longer have to spend a long time in meetings to justify the usage of data and analytics; we can simply dive right into the subject now.
Nevertheless, I still have a strong allergic reaction to buzzwords, like I do to syrupy pop songs of which I tire easily. The main reason, other than I just get sick of hearing them, is because buzzwords lead to kingdom-come-level promises. Overpromises lead to overinvestments, which lead to equally big disappointments (try showing decent ROI on overinvestments), which inevitably lead to finger-pointing and blame-games. That is why I, over time, tried to isolate the beneficial elements of Big Data, and attempted to put different spins and labels on it (refer to “Big Data Must Get Smaller” and “Smart Data, not Big Data”). After all, I am a believer in data and analytics for real-life (i.e., not theoretical) applications, and I want decision-makers and marketers to succeed. I want to find ways to make money with data, whatever you name that activity.
Now I see that the word “Personalization” is becoming the hot topic in conference circuits and the blogosphere. More and more, that word is uttered even during the first encounter with a potential client. Signs are everywhere that it is about to be “the” buzzword in the marketing community.
And I welcome it. Through this series, I have been repeating that the key goals of analytical activities for marketers, regardless of employed channels, should be:
- Knowing whom to contact, and
- Knowing what to offer through what channel, if a customer or prospect is indeed to be contacted.
An amazing amount of data that became available to marketers led to over-communication to an “everyone, all the time” level, and the response rate of any marketing endeavor cannot be sustained that way. Out of desperation, some marketers actually “increase” the contact frequency to maintain the revenue level, and some already have reached a “six times a week, per target” level. What are they going to do after reaching seven times per week? What then? Invent a new day, like Ringo Starr blurted out with “8 days a week”? Spamming more surely isn’t the way out.
Some of my colleagues ask me if we should just take a leap of faith that personalization is the key to the future of marketing, as there aren’t many – there are only few – good success stories about it yet. My answer is to look at all these marketing messages from the consumer’s point of view. Aren’t you completely sick of this barrage of irrelevant pushes, even from so-called reputable retailers? Wouldn’t you pay more attention to something that is more relevant to you, that resonates with you over countless inept and, at times, completely annoying messages? When we show a group photo to anyone, most people check themselves in the picture first. How do “I” look in it? Let’s face it, everyone cares about themselves first, and we are conditioned to pick out anything about us through all kinds of noises.
That is why I believe that this personalization is the future of marketing. In the age of information overload, it is the customers who are picking and choosing messages that are relevant to them, not the other way around. Everyone is exposed to at least five to six types of screens every day. And with new inventions, the noise level will certainly increase. We are no longer living in the world where marketers can just push the products and services according to their priorities. Instead, consumers are ranking products and services. Traditional “push”-type endeavors still have their place in marketing. But in the future, “every” channel will be a 1-to-1 medium, and the consumers will be in full control, choosing what they want to see and mercilessly ignoring irrelevant messages. Marketers must try their best to comply to that demand and show consumers what they may like to see, using all available data and statistical techniques. And the marketers do that right will move ahead. But only if they do it right (refer to “Personalization is about the Person”).
The road to proper personalization is a long and winding one. It starts with the data, of course, as we need to decide “who gets what message” based on them. Various technologies must be employed to display different versions of messages through multiple channels individually, still maintaining consistency. Multiple versions of copies should be written and new stack of creatives must be prepared. Collected data should be refined to be used in such personalization engines, as raw data can only do so much, even with very expensive toolsets. If required data are not explicit enough, or worse, not available at all, we will need to calculate the propensity of certain desired behaviors or consumer characteristics - as in, “not sure if the target is a health-conscious young parent for certain, but he surely looks like one.” As I stated in my previous columns, explicit data are hard to come by, even in the age of Big Data, and we all must make the most of what we get to have. No customer will wait until you have the perfect set of data.
Like in any field, may it be a musical field or martial arts, there are virtuosos (or “virtuosi”?) and grand-masters, then there are mediocre talents and complete novices. In data and analytics such levels exist, as well. Not all analysts or data scientists are on the same level, though I often argue that an unexceptional statistical model is still better than someone’s gut feeling. For end-to-end marketing executions, things get more complicated, as many different types of technologies and skills, as well as overall vision, must work harmoniously to achieve goals. Unfortunately, I often see marketers who still don’t believe in the effectiveness of advanced analytics because they “think” that they had a bad experience with it. But is it fair to dismiss time-tested methods, when many other factors could have gone wrong?
In the interest of not killing the idea of “Personalization” due to unfavorable results from rudimentary trials, allow me to share the “10 stages of personalization efforts” from a data, analytics and technology point of view (i.e., marketing creative is not considered here):
- Not even considering personalization yet. They still think that spraying the same HTML to everyone is alright, as long as the process runs smoothly.
- Personalization is considered, but they do not know where to start.
- Identified basic steps toward personalization, but they do not have specific data or a technology roadmap.
- Created the data roadmap, but they did not start thorough data inventory.
- Identified required data sources, but datasets are not cleaned up or consolidated for 360-degree view of customers (a must-have in personalization).
- Datasets are ready for personalization, but only with “known” (or explicit) data; statistical modeling to fill in the gaps is not considered yet.
- Tested personalization engines through major marketing channels of choice, employing collected “known” (or explicit) data.
- Creating “personas” (or implicit data) using statistical techniques with available data, filling in the gaps with statistical models (an ongoing effort).
- Personalizing most messages and offers through every touchpoint, employing explicit data (known data) and implicit/inferred data (in forms of personas).
- Collecting and utilizing results data to enhance model-based personas and personalization engines continuously, leading to automation.
So, at what stage is your organization? Are supporting datasets previously locked in channel silos merged together to form a customer-centric view? Or are you just plugging transaction or event-level data into some personalization software with a fancy name and a high price tag? Are you personalizing only sometimes through some channels to some people who happened to volunteer – explicitly or implicitly – some of their information to you, or are you doing it for most people, most times, through most channels? The differences are huge. Unfortunately, too many marketers are just personally annoying customers in the name of personalization, and most don’t even do that consistently.
I understand that not all marketing organizations have to achieve ninth-degree black belts in personalization, as from company to company, business models, channel usage, success metrics, budget limitations and available data are undeniably different. Nevertheless, I dare to say that personalization will be more important for the survival of most businesses, as companies that are better at it are visibly leaping ahead. Look at the ways that some big name retailers are doing it from a consumer’s perspective; they are clearly not operating under the old paradigm of “marketers push, consumers respond.” Even when committed to the concept, before any organization gets into the thick of things, decision-makers must set the data and technology roadmap first. The order of operation is important here, and it would be easier to prove the worthiness of the endeavor in baby steps, too. Dismissing the whole idea after trying a few rudimentary steps out of order would be a real shame.
Like any guru would say, awareness is the first step toward improvement. Understanding how far one must go is at the core of any learning process. Isn’t that what Master Yoda tried to teach a young Jedi named Luke Skywalker on that swampy planet of Dagobah?
Stephen H. Yu is a world-class database marketer. He has a proven track record in comprehensive strategic planning and tactical execution, effectively bridging the gap between the marketing and technology world with a balanced view obtained from more than 30 years of experience in best practices of database marketing. Currently, Yu is president and chief consultant at Willow Data Strategy. Previously, he was the head of analytics and insights at eClerx, and VP, Data Strategy & Analytics at Infogroup. Prior to that, Yu was the founding CTO of I-Behavior Inc., which pioneered the use of SKU-level behavioral data. “As a long-time data player with plenty of battle experiences, I would like to share my thoughts and knowledge that I obtained from being a bridge person between the marketing world and the technology world. In the end, data and analytics are just tools for decision-makers; let’s think about what we should be (or shouldn’t be) doing with them first. And the tools must be wielded properly to meet the goals, so let me share some useful tricks in database design, data refinement process and analytics.” Reach him at email@example.com.