Analytics Providers Should Know the Audience
I know for a fact that not everyone is a fan of analytics. Sure, it is a cool buzzword now, but who really likes to look at all those numbers? Even when they are presented in nice charts and graphs, numbers and figures often induce headaches for many. OK, that bubble chart on the spiffy dashboard surely seems nice, but what are those five dimensions again?
Yet we, the geeky types, march on as if every decision-maker must appreciate and enjoy information overload. But I have been saying that “Big Data must get smaller” for a good reason. Basically, most people like simple answers, without even the remote possibility of having to go through lots of numbers; unless of course, they are about sports statistics.
There is a famous book called “Les Misérables” by Victor Hugo. Anyone who was raised in the so-called civilized world must have at least heard about the book, or have actually read or seen versions of it. But the way people consume that classic book (too sappy for my taste, by the way), can be broken down into the following groups:
• Appreciate all 1,900 pages of the original “French” version
• Read the translated “full” version (in English, but still lots of words)
• Skim through CliffsNotes or Wikipedia
• Read the comic book or graphic novel version (yeah, pictures!)
• Attend the Broadway musical (no time to read anything – even with pictures – but willing and able to drop more than $100 per head for it while visiting New York City)
• Watch a TV series (requires a bit more commitment than watching a movie)
• Watch a 3-hour movie (though there is no good movie version of it yet)
• Scan through a one-paragraph summary of the movie
• Listen to someone else’s summary in a few sentences
The final version would be something like, “Yeah, this poor French guy named Jean Valjean stole a loaf of bread and goes through lifetime of hell for it some hundred years ago.” There, “one” sentence. Lesson? “Don’t steal!”
And in the world of analytics, there are similar levels of data consumption:
1. So-called “Data Scientists” who can import any data file, manipulate files any way they want, create reports, charts and graphs with the latest visualization toolsets at will, build statistical models with fluent command of SAS or R, apply the knowledge to all available records in accessible databases, translate findings into plain English, come up with next steps, make business recommendations, and actually sell all these activities for money. Basically super-duper data geeks combined with savvy business acumen. Clearly the top of the food chain, but there aren’t many such tigers in the jungle.
2. Data analysts, statisticians, program/database developers, research analysts or business analysts who are really good at some parts of the long list in No. 1.
3. People who understand ins and outs of data science and/or database marketing and its ecosystem, know whom to call in for projects, and lead the team of scientists to achieve business goals.
Among them are those who aren’t afraid of a sick amount of numbers and reports and can actually derive meanings out of them.
Somewhat like the above, but those who want numbers to be presented in smaller bite-sizes and graphical formats to draw business conclusions fast.
4. People who care more about the colors presented in the dashboard more than the content. (As in “Eew, who uses cyan in the pie chart?”)
5. People who think that analytics is a necessary evil and the actual analytical work is reserved for bona-fide geeks. They may actually feel proud that they are the ones who order those geeks around.
6. Analytics? Who cares? No numbers or stats can top good old-fashioned gut feelings. Wherever that “gut” may reside in a human body (not in the brain, for sure).
Now, here is the major dilemma for people who make living by analyzing big, small, clean and messy data: Who is the audience? How do we even start a conversation about models and advanced analytics, when we start losing your audience at “hello”? How would we evangelize the benefits of analytics in the business world when the majority think that they are not good at math, but are afraid to admit it?
Like in any business setting, one must study the audience first. The business goals may be similar, and the amount of available data and toolsets may be the same, but the way we communicate with the end-users of data and science must be adjusted to the level of the audience, not the data scientist’s level of understanding. Besides, even fellow data geeks may have no patience for all the stories about how it all worked out. Who cares about your 12th attempt that did NOT work? Absolutely no one. You are not dealing with your college professor. Just get to the point, pronto. Let’s get to the business and make some money. And if the presenter doesn’t know the audience? That would be like the Blues Brothers in a Country and Western bar with a chicken wire fence. Good luck with playing that “Rawhide” all night.
For that reason, we the data geeks may have to get out of the habit of presenting the numbers according to the order of operation. We may need to forget about the levels of analytics starting with BI reporting, descriptive analytics, predictive analytics, and ending with prescriptive analytics, whatever that may be.
Or, maybe that word “prescriptive” is onto something. Maybe, people who understand the data and analytics should act more like doctors. Doctors do not show up with prescriptions before they consult their patients. (Some may do that, but that’s a whole new subject). They listen to the patient and examine the symptoms through all kinds of tests before they actually prescribe solutions. Sometimes, they may even admit that they are not the specialists in the field and they call in a new doctor. If their agenda is to push specific drugs or procedures, well then, they just demoted themselves to the level of drug-pushers.
Unfortunately, there are too many pushers in the data and analytics field. Too many make false promises, and too many misguide the masses with one-size-fits all solution packages. I am sorry to inform the marketers that such things never work well, regardless of the price tag. Data solution is about the best combination of available data, toolsets and technologies, not about finding the silver bullet and paying for it.
Analytics must be about solving business problems using any type of technology or techniques necessary, advanced or otherwise. Sometimes all we need is a simple series of business intelligence reports or profiles. Sometimes we may need statistical models to boost the conversion rate and reduce churn rate. Sometimes we may need econometrics modeling to optimize channel, product offering or even marketing budgets. Sometimes we need all of the above, but not in the order of evolution of the analytics, but in the order of client needs. And data geeks must never forget that this whole thing is about the business, not about technology. In other words, re-platforming the data environment into something that sounds cool in the geek circle should never be the goal on its own. Let’s not forget that sometimes all we need to prescribe are two pills of Ibuprofen, not a drill bit through the patient’s skull.
And when the data players deliver the outcome, develop some bedside manners, please. Not everyone enjoys going through the data journey, step-by-step, recounting every bullet. We need to be able to back up the claims with real figures for data enthusiasts, but don’t assume that everyone will want to start the conversation with gory details about assumptions, methodologies and 30-page reports.
One tip? Start with an executive summary in fewer than five bullet points (three are better), and present the next steps in fewer than five bullet points, as well. Imagine the presentation is cut down to 15 minutes without warning. What will your audience retain?
• If the presentation deck is more than 10 pages long, include the details as part of an appendix.
• If any page has more than five bullet points, cut it down.
• If a bullet point looks like a dissertation, cut it down to five to eight words.
Long presentation decks are mind-killers and they stop good conversations. A similar philosophy applies to dashboard design, too: If it looks too much like that of a Boeing jet, try to make it more like an Audi’s dashboard. Most times, less is more — way more. If you doubt that, think about how you consumed “Les Misérables,” unless of course you majored in French literature.
Finally, the users must be reminded that there is no silver bullet in data and analytics, as well. Be aware of the ones who claim that their product and solution are all you ever need. Unless, of course, the machine can do all the things that I listed in Point No. 1 above, all automatically, all in real-time. That will be the day, won’t it? But that is again a whole new subject.
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 firstname.lastname@example.org.