Common sense analytics - the analytics we need

Go to enough technology conferences and you’ll inevitably encounter the amazing motivational speaker. You know the one that climbed Mt. Everest on a unicycle while juggling four chainsaws. From this amazing story, you are supposed to extrapolate that you and your firm can do the similarly amazing things with the vendor's software.

There’s a lot of aspiration pouring out these days regarding analytics and, like the unicyclist to Mt. Everest, it will never happen for most firms. Why? Most firms are still living with kludged technology that delivers data with passable but not extraordinary results. Somehow, the same people that delivered this are now supposed to turn pigs’ ears into silk purses with the magical help of analytics software and maybe some big data thrown in, too. Call me skeptical, but I don’t think it’s always going to work.

Many companies may be overlooking their real analytics opportunity: common sense analytics. These are the first generation analytics they ought to be implementing instead of the shoot-for-the-moon stuff of motivational dreams. Here are some examples of the common sense analytics companies should already be doing (but, sadly, some firms still aren’t):

• Look before you promote – Before a company promotes someone to a highly visible, global executive role, shouldn’t the firm look for evidence that this person likes to travel and do public speaking? A quick glance at the person’s Facebook profile might shed a huge light on the person and whether they’re a private, shy person or a xenophobe who will not travel.
• Pay attention to customer history – I had a phone plan with a major cellular carrier for 10 years. I kicked them to the curb last month for a couple of reasons but the final straw occurred when I returned from Europe in late November. No phone on this plan had ever, in the 10 years we had service with this carrier, had a data plan. Never. While in Europe, one of the phones was set to use data services with a European cell carrier. Upon return to the U.S., the data option was still on and in minutes some 20 mb of email was downloaded to the phone. This triggered an $21 additional charge to the monthly bill. After calling the US cell carrier to see how to prevent this, the carrier offered no suggestions and left me to stew. If an analytic app had seen that we never had a data plan and had not downloaded anything in 10 years, couldn’t it have proactively stopped the downloading until I would have authorized it? No, that would make too much sense.
• Where did our customers go? – I’ve dropped my main airline carrier for long periods of time. I’ve gone six month without using the carrier and I do so when the carrier has failed to deliver appropriate service or has blown me off one time too many. Interestingly, I’ve never once gotten an email or a call from this airline. Had I died or changed jobs? Did I switch carriers? Why?    No, this simple data element wouldn’t be hard to capture if only airlines had an analytic app that could access something like a frequent flyer system that maintains a rich history of a flyer’s patterns. Hmmmm, I wonder if any airlines might have a frequent flyer system whose data they could analyze?
• Check before you offer – It really sets me off when Marketers think I might need something before they’ve done the slightest checking to see if it is even relevant. Why my broadband provider wants to sell me increased download speed when we’ve never needed it, is perplexing. Why some hotel chain thinks I want a timeshare condo in Orlando when I’ve never owned a timeshare and would never vacation in Orlando, is perplexing.  Why are car dealers trying to sell an elderly relative who hasn’t been able to drive for many years a new car/truck, is perplexing. When mainstream businesses make this mistake, they come across as seamy as Internet email spammers.     

What businesses need are common sense analytics – not the kind that require triangulating information from weather satellites, cell phones, third party credit databases and vast feeds from every participant in the company’s supply chain. No, let’s see some real progress with common sense analytics.  What are the guidelines for developing these analytics?

• Before analytics and IT, how did businesses and management know that….? – When you ask this question of business people, you get some great, easy and inexpensive analytic application ideas. For example, how did people know a competitor was failing years ago? They’d see a flood of resumes hitting their inbox. They would notice desperation pricing by a competitor or would hear of quality or delivery problems from this company. How did executives sense a key employee was about to leave? Was it because this person, usually a paragon of attendance, was now suddenly out ‘sick’ a lot? Was it because this person has exercised all of their stock options? The clues are already present for most firms and many problems. You just have to ask one of the older workers how they solved the problem back when.

• Realize that great analytics will require feeds from social and non-social content sources – Now, for you Facebook fans, not everyone in the world uses Facebook. Some people aren’t on LinkedIn or Twitter. Some have accounts but don’t really use them much, if at all. So, relying on these data sources exclusively is likely a mistake. If you want powerful analytics, supplement the social data with other data sources like: transaction data, interviews, phone calls, etc. Sometimes, the old methods can provide additional insights as well as insights into certain demographic groups that may be under-represented in a particular new media source.
• Beware of subversion in social and other data – Teenagers have been misreporting their sexual activity levels to researchers for decades. Don’t worry so much as to why they lie, just know that they do. People leave all kinds of bogus data out on the Internet. Adults on dating sites may provide some of the least reliable information of all. Apparently, there’s a cottage industry of people having long-term dating relationships with non-existent people via social media (e.g., Catfishing). So ask yourself, is it worth targeting a non-existent person for a sales promotion? I rarely provide the entire truth on surveys as I don’t want certain data to come back and bite me as part of an identity theft gambit.  So, how good can analytics be if based on such corrupt data? This is why old and new kinds of data and analytics must be used as well as a healthy heaping of common sense. Verify everything before commiting your firm to a course of action.

Seriously, the best analytics are solving problems businesses have always had. It’s just that now (with low cost analytics, in-memory computing power/speed and access to big data sources), businesses of all sizes can get more accurate answers, more frequently and at lower cost.  These technical innovations are impressive yet they shouldn’t be viewed in isolation. Common sense is still needed – and – from where I sit, businesses need a lot of it.