3 Lessons CMOs Take Away From IT's Flawed Approach to Big Data

There's no question that data analytics are playing an increasingly important role for marketing departments, but many marketers aren't getting what they need from the CIO. If IT can't deliver, CIOs should expect CMOs to go around them.

It's commonly accepted among marketers that data-driven marketing powered by big data analytics is the wave of the future. That has led Gartner to predict that by 2017, CMOs will spend more on IT than CIOs. Others have suggested that the CMO will become the CIO's biggest customer.

Just how that relationship will shake out depends on how CIOs approach the problem of big data and business intelligence (BI). Jennifer Zeszut, former CEO and co-founder of innovative social media monitoring specialist Scout Labs (acquired by Lithium Technologies in 2010), and current CEO and co-founder of Beckon, a software-as-a-service (SaaS) offering for gaining insight from marketing data, says most IT departments have a flawed approach to big data and BI that forces smart CMOs to seek alternatives.

"You keep giving me reports that in no way resemble what I think about my business, what I care about in my business," Zeszut says, speaking of IT from marketing's perspective. "Yes, you give me a million things, but I never look at them."

Zeszut says that she's been in countless meetings in which marketing has been enthusiastic about a new approach or offering until it becomes clear that IT will have to get involved.

"If something has to go through that bottleneck, then honestly, it's not worth doing," she says. "That's why there's a shadow IT organization growing up under marketing ops."

IT's 3 Flawed Approaches to Big Data and BI

Zeszut sums up IT's three flawed approaches to big data and BI as follows:

  1. IT tackles problems from the bottom up. She says IT's plan seems to be: Gather up a big pile of whatever data we can most easily get our hands on, wait for someone to ask a question, then query the database. Business intelligence is essentially an after-the-fact exploration of data.
  2. IT typically pulls in data that's easy to pull in, rather than pulling in what matters. "There's an old joke about a guy taking a nighttime walk who sees another person searching under a lamppost," Zeszut says. "The first guy asks, 'What are you looking for?' 'I lost my keys,' the searcher replies. 'Oh, you lost them right around here?' 'No, I lost them over there in the dark bushes, but the light is so much better here.'

    That, Zeszut says, is what IT is doing when it pulls in only the easiest data. It's true that some of the data that matters most is easy to get (Salesforce data via an API, for instance), but most of it is hard to get. For instance, reams of agency data comes in PDF format. Then there's PR data and all the data and planning docs in PowerPoint and Excel that hold the keys to marketing performance calibration. That data isn't easy to access, so it's probably not coming in to IT's data warehouse any time soon.

  3. IT continues to recommend big, lengthy and expensive data infrastructure projects. When IT presents its plans, they usually conclude with something like, "And then in 2017 we roll out reporting &." Translation: We need to do years of infrastructure work, and after that you might be able to see something.

Those three approaches represent IT's instinctual response to a problem, Zeszut says: Build a massive infrastructure, vacuum up everything, start with what's easiest. But the information marketers need to gain better insights is often locked away in data that's difficult to pull, and they need access to it ASAP, not six months to a year or more down the road.

"For all of the above reasons, marketers who want to understand the business impact of their marketing can't just leave everything to IT or the analytics department and say, 'See what you can find in there,'" Zeszut says. "Are analysts going to serendipitously sift out key insights for optimizing your marketing? Not likely."

Instead, Zeszut says, marketers are learning that to benefit from big data, they must take control and become "the storytellers of their own success."

Marketing is hungry for data, she says. But unlike other business functions, its access to that data is hobbled.

"Think about if we stood for that on the finance side," she says. "Every time there's a purchase order or expense, please put all of your receipts in the slot of this warehouse door. Wow! We have all of the receipts in one place. Now if the CEO wants to know how much we spent on software, hold that thought. We're going to send someone in to sift through all the receipts and in six weeks we'll have an answer for you."

"But that's not how finance works. Finance tags, structures and organizes data on the way in. Finance knows at a moment's notice where we are on anything. We've chosen to think about the kind of reports that we need at our fingertips on the way out, so we tag everything on the way in. We have instant visibility into what's going on."

Making the IT Function Relevant to Marketing

CIOs can help CMOs achieve that level of visibility if they're willing to change their approach, Zeszut says. IT may be worried about security and infinite flexibility, but marketing wants answers to its questions, ASAP.

"Start with the really critical business question that the client, the marketer, has to know the answer to," she says. "If you start there, it makes it super clear what data sources matter and which don't. Not only does it tell you which data matters, it also tells you how to pull that data in and how to structure it."

For instance, marketers spend much of their time talking about the marketing funnel, with awareness-generating activities at the top and purchase at the bottom. Different marketing activities target different portions of the funnel. A classic marketing question is: how much am I spending at the top of the funnel versus the bottom of the funnel, and is it the right allocation?

Top of the funnel activities include things like advertisements on television, billboards, radio and in magazines. That's data that's difficult for IT to access. More importantly, if there's nothing in the data warehouse labeled TOFU (top of the funnel), MOFU (middle of the funnel) or BOFU (bottom up the funnel), answering that question will be an analytics dead-end. Unless the data is appropriately tagged as it comes in, BI analysts are going to be stuck conducting difficult and extremely time-consuming data forensics.

"However, if you know you need that answer, there are ways to make sure you've got it at your fingertips," Zeszut says. "You could add a new field to your purchase order form so your team can tag expenses by primary intent in your funnel."

Thor Olavsrud covers IT Security, Big Data, Open Source, Microsoft Tools and Servers for CIO.com. Follow Thor on Twitter @ThorOlavsrud. Follow everything from CIO.com on Twitter @CIOonline, Facebook, Google + and LinkedIn. Email Thor at tolavsrud@cio.com

Recommended
Join the discussion
Be the first to comment on this article. Our Commenting Policies