SXSW 2013: Small Data in a World of Big Data

KPI-originalMore than ever, modern marketers are using data to drive decisions about strategies and tactics. Clients are asking us to back up “what we think” with real numbers about “what we know.”

The rise of “big data” is a hot topic at this year’s SXSW conference. Thought leaders are sharing their insights on how to interpret and act upon this growing pool of information. Pay attention, since this stuff matters.

But at the same time, we run the risk of ignoring “small data.” That is, if the numbers we have do not hit a certain threshold of massiveness, should we throw it away? Of course not; let’s discuss why.

Aggregate Shmaggregate

A few years ago, economist Steven D. Levitt and writer Stephen J. Dubner released an important book called Freakonomics. This book addressed some interesting ways data can lead to improper and unfortunate conclusions. Specifically, big data can make you think that 1 + 1 = 3. (Note: It doesn’t. I checked.)

Freakonomics revealed that sometimes big data, when combined with unrelated external events, can confuse even the smartest, most experienced researchers. Add to that the idea of “unintended consequences” and things can get downright confusing for people who put all of their faith in big data.

There’s nothing technically wrong with big data. I love big data. Send it over and I will have my big-data brains dissect it. Big data = good.

The problem is twofold:

1. Big data is overwhelming to many people. They draw incomplete conclusions based on their limited view of the numbers. It takes a well-trained, experienced data professional to extract smart marketing insights from gigantic data sets. When a well-intentioned but inexperienced person analyzes data, he or she can easily misinterpret the insights.

For example, if tons of people are landing on a particular page on your website and then leaving immediately, you might conclude they are “bouncing.” We hear this a lot.

But what if they simply found what they needed and left? What if this bounce was really a conversion?

This leads directly to the second problem…

2. Weak KPI mapping. Key performance indicators (KPIs) are designed to help you understand the relationship between your content and the target user. Specifically, how well your content satisfies their needs and moves them along your relationship continuum.

All the data you collect means nothing until you set KPIs and other goals. KPIs give you context to the data. Without this context, you will get data for data’s sake. And really, who wants that?

KPIs need to be an agreed-upon measurement that guides content creation, traffic drivers, and analysis. (That’s it. Please reread that sentence aloud to the whole class.)

And Now, Small Data

So, what’s the takeaway? Well, it’s about how you should be interpreting your data, both big and small. It’s a new mindset that you need to bring to your team, who are probably building a bonfire and chanting, “big data, big data, big data” the week after SXSW.

Big data matters. I completely agree. It’s just not as simple as it seems. You can’t just look at a giant bucket of data and make a snap conclusion (e.g., kittens are popular, hence we need kitties on our website to sell our product).

But small data matters too. Yes, you can learn a lot about the time of day your site is visited. And yes, it can be extremely significant to chart this over a two-year period.

Small data, however, can tell you different things about your target. It can tell you if the user is getting information that influences their decision or engages them  in your resources. Small data can tell you if your content is properly linked within your website and expanded brand footprint.

For example, a “thank you” page at the end of a transaction page may be a KPI. If you’re only looking at the big-data picture, you may not really understand what people are doing to reach this page.

You may not even know how they got to your site to get to this KPI. Search? Banner? Email? Sure, you know how they got to your website, but do you know which channel delivered the user who converted? In too many cases, the answer is “no.”

Everyone from the copywriters to the designers need to know your KPIs. Your SEO team and your media-buying team, too. Everyone needs to know your KPIs, most of which are not part of your “big data.” KPIs are often part of your “small data.”

Looking forward, make this year the one that everyone is measuring data consistently. Understand and optimize your website—and every other tactic in your campaign—according to big data AND small data.

And skip the kittens.

Check out OCHWW’s other SXSW 2013 blog posts:

SXSW 2013: How Zombies Are Helping Us Get Fit

SXSW 2013: BIG Data and Personal Technology at SXSW

SXSW 2013: The Mobile Healthcare Revolution

SXSW 2013: Bad Behavior – the Saga of SXSW

SXSW 2013: Empty Information Calories

Questions? Comments? You can contact the author directly at
Please allow 24 hours for response.

This entry was posted in Analytics, Content Strategy, Data, Healthcare Communications, SXSW Interactive and tagged , , , , , , . Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. Matt Balogh
    Posted March 8, 2013 at 4:07 pm | Permalink

    Great post Buddy! I’m working on a personal initiative for Lacey Elks Run for the Heroes for which I’m using targeted campaigns because of my [very] limited budget. The interesting thing is what I’m learning about the segments I thought I knew. Highly targeted campaigns may appear to have great ROI, but there’s also an opportunity cost of those you’re not targeting. Big data can tell me what’s happening, but it can’t tell me what’s not happening (who I’m not showing my ad to). Example: my Mom doesn’t run, but if I don’t show her the ad she can’t tell my Dad about it! That’s an important distinction which can truly make the difference!

  2. Buddy Scalera from Ogilvy CommonHealth Interactive Marketing - NJ, North America
    Posted March 8, 2013 at 5:28 pm | Permalink


    Agreed. At the end of the day, we need to use both the macro and the micro views of data. We can learn from both sets of data, if we just take the time to analyze them appropriately. Hopefully this gets people to plan out their KPIs before they start creating and analyzing.