Our new GreenBook Directory site is live!
Your #1 strategic guide to consumer insights.
Qualtrics X4 The Experience Mgmt Summit
Brand & Retailer tickets for all IIeX events now start at just $99! Get or give one today!

Are You Burning Away Your Data Fuel?

As data-driven approaches take organizations farther than ever before, the need for well-designed databases is more relevant than ever. If competitors are able to harness the power in their data fuel, then they will be the ones to go places with it, leaving us in the dust.


By Patricio Pagani

The ‘data is the new oil’ metaphor is overused these days. And it’s not particularly apt. After all, we aren’t drowning in an excess of fossil fuels, are we?

But if data is the fuel that powers your business, then we have a problem.

We’re setting fire to our fuel with the inefficiency of 19th century colonials in a Kauri forest.

If your data is oil, a well-designed database is your motor

jet engine 08 - graphic

Just like setting fire to our fuel can give us short-term warmth, a primitive approach with data management gives us haphazard results. We’re not seeing the fullest picture which our data could give us.

Five tell-tale signs that your organization is wasting its data fuel:

  1. You sit through meetings where oodles of meaningless data sprawls over endless PowerPoint bar charts.
  2. When the somnolent audience finally speaks up, it’s to question something which looks wrong.
  3. The speaker can’t explain the mistake.
  4. Information-sharing initiatives are launched, the quietly fade away…
  5. Management prefers to make decisions based on instinct, because using data-driven insight is just too hard.

For most of us, the word ‘database’ conjures images of the clunky interface of early versions of MS Access, mysterious acronyms and lines and lines of dull-as-dishwater raw data. Hearing the word makes people either go into a panic that they don’t have database skills, or switch off because it sounds out-moded.

Perhaps a better way of thinking of databases is as a well-organized storage facility or knowledge bank: an organized structure in which to store all that valuable data.

It’s curated by skilled data managers. You can dip into it whenever you want, and the data is ready to use when you do. In this ideal world, you won’t have to repeat a huge research study because you couldn’t access the study that your predecessor (or somebody in a different business unit) run. In this ideal world, all the research studies that your company has purchased over the years are available at your fingertips. And before you decide to ask to burn new fuel, you will definitely check that the tank is empty,

Locked in silos


Most large businesses have several agencies supplying them with consumer insight data, often in many styles. But marketers need to look at the bigger picture. Keeping your data-fuel in silos (be it PowerPoint decks of study results or even an Excel dashboard with prettily-formatted tables) doesn’t help marketers drive brands forward. It shouldn’t really matter to marketers where the insights come from. They are agency-neutral. By bringing the data together into a single knowledge bank, you can:

        Mine and compare consumer insights across studies from one country or many countries

        Find new patterns and relationships in your data

        Get more value from your existing research and save money because you don’t need to repeat old questions just because they aren’t accessible.

 Speak a common language


But in order to be able to extract all that value from your investment in consumer insight, the first step must be to plan the way you want to look at the answers. We call this the ‘data architecture’. The best architecture for your data is one that is aligned with your business.

It’s tempting to organize data by source, or by the internal team who commissioned the project. But what if we organized it by the kinds of questions which marketers ask? For example, ‘Where are our greatest opportunity segments?’ or ‘What impact do our campaigns have on awareness?’.

Don’t settle for ‘researchey’ labels like ‘Grouped socio-economic class of HH’ on your variables. Take the time to align them with internal language. The outputs will look less intimidating to the non-researchers you want to engage.

Even more importantly, align the labels of common variables between studies. For example, you have a weekly behavior tracker and a monthly brand tracker. Wouldn’t it be great to overlay brand performance on purchase behavior? Net up the weeks into a new month variable in your behavior tracker in order to easily look at the two studies side-by-side.

Harness the power


As data-driven approaches take organizations farther than ever before, the need for well-designed databases is more relevant than ever. If competitors are able to harness the power in their data fuel, then they will be the ones to go places with it, leaving us in the dust.

Let’s stop being gas-guzzlers, and start being disciplined data consumers. We’ll cover more distance when we use our fuel in a finely-tuned motor.

Please share...

3 responses to “Are You Burning Away Your Data Fuel?

  1. British Prime Minister Benjamin Disraeli (1804–1881): “There are three kinds of lies: lies, damned lies, and statistics.” The real truth is that data. like statistics, can be many things, depending on the way you view it.

    The real value of any data is consistency within and beyond the context in which it appears. Media, in particular, tends to generalize data points and force them into arguments while marketers often use data to differentiate. In traditional market research we created finite universes that were specifically designed to tasks. Those findings fell within a specific margin of error and were rarely applied beyond the study at hand. Contrast that with the way data is managed today. In many cases we are using infinite sources of data collected from multiple channels which are screened, scraped and often collected through interactive transactions. There are no models or methods of normalization that can account for the levels of variability that exist within the data and even if there were, they couldn’t provide answers in the time frames needed for mobile transaction management. That’s why Google questions and quick polls are rapidly becoming a major part of the researcher’s toolbox. Census measurements are better indicators within a tight time frame and varied audience.

    The key to disparate data management is finding consistent patterns in the sources that exist across both a time and geographic variable. The programs have to be sophisticated enough to pick-up on emerging variables and descending trends. What makes it difficult is that aggregators such as segments and personas remain fragmented, particularly when behavioral metrics are added, such that the programs cannot identify large blocks of opportunity. Segments, segmentations and personas depend on the ability to group attributes and behaviors. In an interactive society, behaviors are influenced by many factors and there is no stimulus to measure against. We are in effect, asking for an elasticity model but we’re not providing any static variables and an infinite number of variables that are interdependent with fluctuations in dominance.

    Why not instead try to focus and understand the appeal of the brand and its products? Those have finite and discreet attributes. If your data sources identify via census or otherwise that a substantial number of people want a no-kill mousetrap can you not just ask where people want to buy no-kill mousetraps. You can certainly do the product testing with your current audiences. Do we really need to know their motivation or what they plan to do with the mice or even why the have mice? C-level and product managers just wants to sell the mousetraps. That’s why BI is succeeding and MR is retreating. At the end of the day it’s about maximizing sales and profitability. Somewhere along the line we got tangled up in the weeds.

  2. Hey Ellen,
    Thanks so much for your thoughts on the topic. I’m not sure I agree 100% with Mr Disraeli, but I’m a researcher, aren’t we supposed to love stats?
    An interesting discussion we are having with our clients at the moment, and I’ll be presenting a paper at ESOMAR with Microsoft to discuss it is how traditional research (or perceptual research) can be complemented with behavioral data (aka Big data). There are huge debates happening in client organizations left, right and center. Which data is more accurate? And for the reasons you highlight above, is data that can’t be understood without slicing down be trustworthy?
    We also shouldn’t forget about a topic that I don’t think our industry has discussed enough.

    While the main purpose of research -whichever way you do it- may be to provide information so our clients can make better decisions, the same information/metrics are used for C-level compensation plans. The bonus of our client bosses -and the bosses of the bosses- are paid partly according to the results from our research.

    So choosing the right (or wrong) tool, not only would have an impact on the quality of the decision making, it can also easily get some people fired. I’ll hold my thoughts until on how stable/reliable Google and other quick poll tools can be for the purpose of this. Not one solution would fit all, and we’ll soon realize this is true, in the ‘new’ world of market research, same as it has always been.

    My personal view is that in the future, short polls will coexist with behavioral big datasets and be complemented with “traditional’ surveys, only shorter than those we usually conduct today.

  3. I recall having a conversation with an Executive about 8 years ago about the most progressive, innovative and successful companies being those who focus on getting their databases up to speed. Without strong databases you can’t mine for insights, communicate with your customers, or manage your business effectively.

    Thanks again for the great article Patricio!

Join the conversation