The Project Every Retailer Needs And No One Wants: Big Data Marketing Automation

Todd Michaud runs Power Thinking Media, which helps retailers and restaurants tackle the convergence of social, mobile and retail technologies. He spent nine years delivering technology solutions to more than 10,000 retail locations as VP of IT for Focus and Director of Retail Technology for Dunkin' Brands.

Retailers everywhere are finding themselves being hit upside the head with big data. This data is generated by their internal systems, external systems and end customers, and it's growing at exponential rates. Quietly, this trend is going to add a new player to the corporate ranks: the Chief Data Scientist.

Organizationally, these functional experts will challenge the traditional organizational structure. Data scientists will likely enter an organization through the IT group, because that department is most likely to engage their services to help drive value from information mining projects. The challenge comes in the fact that, today, many organizations group IT with finance and accounting teams (likely to utilize the CFO as a buffer between the CEO and CIO).

But keeping the data scientist under the CFO is problematic, simply due to the front-office nature of the role versus the back-office slant to finance and accounting and, in many facets, IT. Over time, I see the data scientists eventually becoming a unique functional area that is broken out from the IT group. Although this may go against the grain of "keeping the geeks together" philosophy of many current organizational structures, I see the data scientists pushing for their own independence. Their work is much different in nature from IT (versus deploying managing systems), so it doesn't truly align there, but it is threatening to the marketing team and the art versus science debate. So with no clear alignment with IT, finance and marketing, it's only logical for data science to become a separate and unique function.

Furthermore, once the executive team realizes the value this role is delivering to the organization and how they own the "secret sauce" of their business operations, it is likely that the CEO will want to keep data scientists in close company. This could eventually lead to a similar relationship to the one they have with the CFO.

The next generation of business process automation will be the operationalization of big data in the form of Marketing Automation. Retailers will focus on implementing tools that consume this information, not for the purpose of generating reports or dashboards but with the intent of taking action.Although marketing may soon have a larger technology budget than the IT team does, marketing chiefs may also find themselves taking a back seat to a mathematician (data scientist) when it comes to reporting marketing ROI to the CEO. Who do you think a company will value more, someone who works with an agency to create a great commercial series or free standing inserts or someone who figures out how to optimize the retailer's relationship with each consumer through advanced algorithms?

I'm not saying that marketing will move from an art form to a science completely or that this change will happen overnight. But I am saying that data scientists will have hard numbers (ROI) to backup their activities, where a marketer will be using Neilson and focus groups. Where do you think the money is going to go?

Retailers are looking at tremendous opportunity to leverage their data for improving sales. Netflix has spent millions of dollars in crowdsourcing the algorithms for its recommendation engine because it understands the power of that data.

Many people scratched their heads when Wal-Mart acquired Kosmix for $300 million and launched Wal-Mart Labs, but, to me, it makes perfect sense. It was a talent acquisition with some nice Intellectual Property that will help the chain launch its need for advance coding, data mining and processing.

This isn't a small challenge. Processing big data is hard (and beyond most retailers' capabilities today). But automating processes based on the information extracted from that big data processing, that's infinitely harder.

To give an example of what I mean, imagine processing every tweet from all of your Twitter followers and determining what actions you need to take with each person in an effort to maximize your relationship with them.

Do you send them an E-mail? A text? Present them with an offer at Point of Purchase? Do you offer them a 10 percent discount? Twenty percent? Thirty? Is it for a specific item or category? Do you offer them a perk instead of a discount? Do you offer something they can pass along to their friends and followers?

You can see from my example, which only talked about one social network, how complex this can get in a hurry. But it also shows how valuable it can be. A recent Forrester report shows that the majority of Groupon purchasers would have gone to the retailer making the offer anyway and paid full price. Ouch. What if you could identify that up front and offer your loyal, would-be customers a smaller cost perk instead of a 50 percent discount for their business?The great thing about science versus art is there are very clean lines of success. Science can't tell you if your new logo is helping your sales over the old logo, but it can tell you that a small tweak to your loyalty algorithm yielded a 0.001 percent positive variance.

Two big things stand in the way of this type of system being put in place by large retailers: money and organization. Let's face it; building a system that can process large amounts of data, and then automate a newly created marketing platform, is not a cheap endeavor.

I doubt that you will find many Chief Data Scientists out there today, but you will in the future. This functional group will likely be separate from the IT team., First, the skills required are highly specialized. And second, if it were to be placed under the CFO (the unfortunate home of many retail IT teams), the program will be doomed, because most financial leaders just don't have the required sales and marketing mindset that this group will need.

Add to that, many CEOs already have frustration and trust issues with their CIOs over their track records for delivering enterprise technology projects. Putting this much of the company's financial resources into a project that only Ivy League math majors truly understand the details of is not something they will take on lightly. In my experience, if the CEO or CFO doesn't understand how it works, they don't really trust it and, many times, will not do it—regardless of how good the business case looks.

The CMO is not likely to be a supporter of the project, because it's "not just about the damn numbers, it's about the feeling." The CIO is not likely to step up to the biggest project of his or her career (which this will easily be) knowing that the results of the project will easily be measured by all—"Creating and defending a business case is for my business partners to do."

So you've got a business transformational initiative that is super expensive and that no one in the organization is really incented or happy about undertaking at the senior levels. Sounds like a winner, right? In many retail organizations it will be easier to just let inertia take hold and not tackle this type of problem.

Companies like Amazon and Netflix have "embraced their data side," because it was a foundational part of their business from the start. But if you've lived by the "run a commercial and see a 3 percent sales increase" model for as long as you can remember, this transition will be a difficult one. It's going to take some serious leadership from the top, with a healthy dose of "You're damn right we are doing this. Get on the bus or start walking" mentality.

But, as I've said before, the retailers who get personal with me first will be the ones who get my business, a bond that will be difficult to break. Do you really want to let the other guys create that bond first?

What do you think? If you disagree (or even, heaven forbid, agree), please comment below or send me a private message. Or check out the Twitter discussion on @todd_michaud.