Big Data For Marketing? It’s All About The Questions

Looking into a heavy topic today: Big Data for Marketing. “Heavy” because it’s sorta philosophical. I think Big Data for Marketing is more about the questions than the answers. More about the insights than the technology. The opportunity lies in defining how Big Data can help you better reach your customers with your messages in a way that gets them to act.

And this starts with simple questions like:

• How did your last campaign perform?
• Which keywords drive conversions?
• What websites did my prospects visit before they came to my sites?
• Which social conversations are spurring the actions that lead to deeper engagement with your brand?

Big Data for Marketing is easy to say but hard to do. So here I will look to a few of the “experts” on how to take advantage of all this data and turn it into real business insights that drive real business value.

Big Data Marketing Is Like Gold Mining

Dean Bedard likens Big Data marketing to gold mining when he says “it’s not enough to know where the precious stuff is. In order to reap the riches, first they have to get it out of the ground and turn it into something they can use.” He outlines four steps to prioritizing the road to mining Big data insights including:

• Setting up a cross-functional marketing and IT team
• Prioritization of the marketing goals Big Data can help you accomplish
• Mapping the data sources to obtain reporting on key metrics supporting the main objectives (KPIs)
• Creating an “agile marketing” implementation roadmap which develops against the highest priority areas to generate the quickest value.

Big Data For Marketing Requires ‘Destination Thinking’

Matt Ariker from McKinsey had similar thoughts when he suggests marketers start their Big Data projects by thinking of the end goal and then working through all the details. This so-called “destination thinking” will help the strategic marketer avoid the traps of many Big Data Marketing projects where the deliverable becomes the end goal itself instead of the business value imagined at the outset.

The Era of Big Testing

Scott Brinker (@ChiefMarTec) suggested in my own future of marketing interview series that we need to encourage more marketing experimentation and “Big testing.” He believes marketing needs to take advantage of both new technology and new talent to start creating hypotheses. Then to use Big Testing to prove them out – right or wrong. For Scott, “the key to scientific marketing is actually the embrace of marketing experimentation as a driver of continuous innovation.”

Marketing Offices As Trading Room Floors

Brian Kardon (@BKardon), CMO of Lattice Engines imagines a day when marketing offices resemble trading desks with screens of real-time data streaming in and marketers yelling “buy, buy.” But he laments, the effort by most marketing departments to develop real-time dashboards are “dwarfed by countless marketing Powerpoints and never-read marketing plans.” Brian understands the power of questions. In his own office he has dashboards that tell him:

• What was the unsubscribe rate from that new campaign?
• Which version of the email performed better – the one with the orange button or the blue button?
• What deals are stuck in the pipeline for more than 60 days?
• Which reps had the lowest win rates this year to date?
• Is Twitter sentiment for our brand trending up or down?
• What % of our customers at Dell are using Lattice right now? How does this compare to yesterday, last week or last month?
• How many unique visitors came to our site today?  Where did they come from?

Big Data Marketing Must Be Illustrative

According to Jon Baron (@TagManJon) the difference between Big Data and Big Data Marketing is found not in the analyses, but in the insights that turn into actions. Jon defines Big Data as “a simplistic term which refers to the automated accumulation and analysis of audience data on a large scale.” But emphasizes that Big Data for marketing must be illustrative of the target audience the marketer is seeking to understand. It must help us to be more efficient or effective at reaching a target audience and getting them to act on the marketing message.

Big Data Marketing Means You Understand Your Customers’ Questions

Finally I’ve always talked about the need to understand the questions that our customers are asking when they do a Google search, visit a website or participate in a social media conversation. I believe every marketing plan should start with an analysis of your companies “share of conversations.” This shows you not only how well known your company is in your solution space, but also how likely your customers are to act on that awareness (RT, comment and ultimately convert to a lead and a customer).

My own 4-step plan to Big Data Marketing Insights:

• Define your customers’ top keywords and topics through search analysis. Starting with a simple look at Google Trends.
• Social analytics to show what % of social conversations include mention of your brand. This is where the data gets really big.
• Insights from your Website analytics will help you understand which keywords and referring domains drive conversions.
• Finally, combine these 3 massive big data sets to help to make decisions on marketing plans, tactics and budget.

What Marketers Are Saying About Big Data

Here is a curation of thoughts from folks I follow. You should also let us know what you think in the comments below.

Steve Peck:

“I believe big data will continue to be talked about and properly utilized by the top 1% of savvy marketers, but for the mainstream marketer, it will continue to be like the mystic marketing leprechaun. In other words, something they will continue to search for and talk about in group settings in hopes of scoring a higher ROI and pot of gold, but have no clue how to actually find, capture or put ‘big data’ to work when they return to their cubical. Not until larger numbers of innovative companies develop turnkey ways to generate easy to understand and actionable insights from big data will it become ‘the year of big data’.

I believe that marketers in 2014 will continue to make inroads to get closer to capturing the big data ‘leprechaun’. Instead, the year will pick up where 2013 left off with continued focus on the needs and wants of the individual reader, and a focus on understanding what qualified customers are most interested in, so that sales and marketing professionals can automate delivery of that relevant content, or product specific information, based on where that prospect is in the buying cycle.”

Eric Wittlake:

“While I generally agree with Steve’s premise, I think we can consider big data in marketing more broadly as well (strange to say, as it isn’t particularly well defined for many marketers either).

Consider the algorithmic learning and optimization a DSP does across billions of impressions with 100’s of data points per impression, all for a single marketer. While it may be a limited application, it is something many marketers are embracing (with mixed results I’ll add, but that isn’t a fault of the premise of big data).

Social media is another space where some of the ideas of big data are being applied, and marketers are making use of them. A question marketers are starting to use this information to answer: What types of conversations are predictive of someone in market? Others are using it to identify emerging trends, like watching for words paired with their brand or category name and identifying emerging pairing and those that are trending up or down (simplified example of far cooler things real data geeks can do with this data).

When we look at it this way, marketers may not be sophisticated yet, but many of the tools they are using are giving them some of the benefits of big data today. Now, instead of pursuing big shiny data infrastructures and software, they need to pursue information or insight they can’t get today using these existing solutions. When that leads to their own big data toolsets, it will be because they really have an application for it.”

Howie Goldfarb:

“Businesses have crunched data forever. I remember in the mid-90’s and I was trained as an ISO9001 auditor and my employer used constant improvement metrics. Our vendors were moving to SAP which was/is a big data platform especially for procurement chains. Then in the mid-2000’s my new employer was using lean manufacturing and working on Six Sigma. Sales has always used Big Data. Marketing too. Just now the tools are less expensive, and data is more widely available to the average person. Google Analytics is Big Data.

So I feel marketing needs to stop with the hype of jargon and focus on the nuts and bolts. We should be talking about how small businesses with limited resources can now have data tools available that previously they had to pay for from the likes of Oracle, SAP, Vocus, etc. I mine data from the media. I crunch numbers with my finance background daily. The key is training people who to look for the data and use it. Marketers tend to pick and choose what data they use (subjective). For example, the average Facebook user engages with each brand page they LIKE less than 1x per year. I didn’t need a database to find that data just math.”

Doug Strickland:

“While the idea of real-time dashboards are very cool indeed, there are a couple of problems with their execution.

First, social marketing dashboards aren’t like automobile dashboards. Better said, we aren’t as good at driving the insights bus as we are the family van. Having data in front of you doesn’t – real time or not – provide the insights and analysis that comes from a little ‘soak’ time. We will get better at understanding the dials and indicators over time, but right now it’s shiny buttons and flying lights and most analysts struggle at differentiating the speedometer from the tach.

Add to that the service layer – or perhaps subtract. We have a lot of trouble getting real or near-real-time data from the currently available constantly changing API’S.

Our organization is working on surfacing the “meaning” more than just the data. Just like there are narrative character archetypes in literature, there are patterns in social behavior that are discernible, measurable, and can be recognized best through automated systems.

While most fumble with keywords and topics, imagine if the dashboard told You what was important by recognizing shifts in topical points of focus, meme transitions and evolutions, inflection point patterns, etc.

When dashboards can teach us to fish…”