“A lot of direct marketing activity is retrospective; it’s about trying to replicate the past. With predictive analytics, you’re being proactive. It can turn the tables,” Simon Kaffel at HSBC said recently in a MarketingWeek article, touching on the excitement many marketers feel about the promise of predictive analytics.
Predictive analytics is the use of data mining, statistics, machine learning, and modeling to analyze current and historical facts to make predictions about future events. Predictive analytics is like cat nip for marketers. It holds the promise of being able to present the right message to the right people at the right time.
But we’ve all experienced marketing that gets it wrong. An ad might cue off of someone’s browsing data but miss that they already bought the product somewhere else. Marketers might have a detailed record of actions taken by potential customers, but completely miss the intent behind those actions. And so, people are inundated with marketing messages that are partly right, but largely wrong.
Marketers have more data at their disposal than ever, but we’re still in the adolescent stage of figuring out how to use it. It’s a struggle to draw the right insights from so many data sources. There’s a large gap between the promise and the reality.
As predictive analytics continue to improve, it will transform how we work. But there’s a massive learning curve. I’d love to hear your thoughts on using predictive analytics for marketing.
Here’s a cartoon I drew three years ago about the promise of Big Data.
jonathan goodacre says
It’s funny this should drop in my inbox as I was doing some predictive work for a client and through my head were going all the problems and issues. So I put lots of corrolaries in my report, but that can be difficult too because then it seems like we don’t have confidence in what we are doing. I suppose it works irs way out as the good stuff works and we get better at using it. It’s a worry though!
Steve Masters says
The thing to bear in mind with big data is that it reflects trends, and trends are not intrinsically linked to needs or desires.
We are programmed to be drawn to a crowd, so if a crowd builds around something pointless, the trend suggests we like it. Take, for example, television. Reality shows have mushroomed in recent years, and they have become more and more risqué. Data would suggest this is what we want more of, but is it really? Is titillation, conflict and the shock factor what we crave or is it simply what we are drawn to?
If the government had used Google’s Flu Trends to stock up on medicine, it would have made far too many drugs.
If movie stats are anything to go by, love of Star Wars outweighs almost everything else, but millions of people, possibly billions, aren’t that interested in science fiction, let alone Star Wars.
I’ve been reading Small Data by Martin Lindstrom, and it is an excellent adjunct to the concept of big data, because it studies motivation and behavioural trends related to society at large.
Woody Savage says
Predictive analytics for marketing could be more of a science if pursued in the right way. Many people “frown” on sharing their “private info” but I would relish sharing MY PROFILE if I thought it would better direct advertising and email messages to my NEEDS rather than via a “shotgun” approach. My first wife who died 12 years ago still gets mail, people vote who are dead or have a driver’s license in some states but no U.S. citizenship. A national registry of everyone born and dead could be a foundation of who is in existence. Then you could at least address “age appropriate” solutions or ads to people. At 72, I don’t need investment advice for retirement as I have been retired for over 3 years! A profile would allow solicitors to at least approach prospects with more meaningful solutions to their true needs.
Ted Vasquez says
Predictive analytics has enormous potential for marketing and really every business function, in every industry. Like ecommerce around the turn of the century, predictive analytics seems to have a certain manifest destiny. Yet, as you point out, there are many examples of its current limitations.
In my experience, leveraging predictive analytics is a journey of continual improvement. Most companies lack the talent, data, infrastructure, and business processes to tackle complex personalization scenarios right away.
Following the illustration, predicting what a customer will want next is one of the most difficult questions to answer correctly. There are simply too many possibilities and the data is often quite misleading. Past purchases would seem helpful but they may have been motivated by gifting, random events, particular fads, etc.
Instead of trying to predict what specific item a customer may want next, companies may want to start with easier questions like will this customer’s next purchase be a gift or a trending item?
With those higher level insights in hand, Marketers have a great opportunity to increase relevancy… while the Insights Team digs deeper…
Christian Gatlin says
Great post!
Predictive analytics definitely has the potential to be very beneficial to every company, and with the customer’s desires continuing to change there definitely is a gap between forecasted/projected wants and reality. It’s imperative for companies to stay updated on new trends and the ever-evolving market; otherwise, we will never be able to figure out how to properly segment and apply big data.
Hopefully when businesses decipher these analytics they will produce an accuracy that will manifest an increase in sales and profitability.
Sharon Green says
I’m very interested in predictive analytics and will be in the workforce long enough to see how it develops over the next 10 to 20 years. I also think we need to be mindful that more data isn’t necessarily better data. It’s also important to test out hypothesis and work out the stroy the data is telling us. I’m hopeful as we becoming better at sifting data and interpreting it, we can enrich experiences and it will impact more on sales, productively, engagement and profitability.