Data-driven focus leads to falling off cliffs
Originally published at www.linkedin.com.
Data is the fashionable thing for marketers at the moment. Whether it’s Big Data, small data, bits, bytes etc. Don’t get me wrong, data is great. It’s been around for yonks. But don’t forget that data is only as useful as the people using it. Don’t believe me? Be lucky you weren’t this data-driven driver: Man follows sat nav to cliff edge.
In fact, it wouldn’t surprise me if in a few years we’ll be reading articles titled “Big Data is Dead”, similar to the current bout of “SEO is dead” content, which has existed ever since SEO has. Even though you know these titles are exaggerated and attention-grabbing it’s hard to look the other way. (My title is a bit like this! However I at least included a link showing it is true)
How not to use data
Scientists have always used data. However they normally start with hypotheses, and use data to prove these. This is the right way to approach data.
So, using data to form a hypothesis seems a bit back to front. For example, some tactics based on “data-driven” methodologies without clear objectives in mind have been shown to miss the mark somewhat:
- Focusing on Facebook likes or Twitter followers, or even worse, a Klout score, as a means of influence.
- Telling your users to share or like your Facebook or Twitter statuses.
- Installing a Hadoop, MongoDB, or NoSQL database without understanding why.
So how should you use data?
You should always start with ideas, and what will work best for you or your client. Experienced marketers will have a good instinct for what will work, and what won’t.
Marketers new to the business will often have fresh ideas that haven’t been tried before which are worth testing. Some will fail, but some will succeed in ways people never imagined.
When data works
There is a place for data where an idea has been tried and tested and put into practice.
- Motorway Variable Speed Limits, as going slower actually increases average speed.
- Amazon’s Killer Email Marketing, using the information you have on customers to suggest other items they may be interested in
- How Netflix Gets Its Movie Suggestions So Right, they started with a belief this would work, and through trial and error — even offering a prize to people who could improve the algorithm — 75% of what Netflix users watch now comes from these recommendations.
The data didn’t come up with these ideas, people did, and they used the data to test their theory before implementing big changes.
Do you need “Big Data” or just “Data”?
Most people don’t need Big Data. They just need a bit of data to start testing a hypothesis. Unless you’re a company like Google, Facebook, or Brandwatch, or a large business like Tesco who all have enormous data sets. Facebook actually still use MySQL. There is a gap between Microsoft Excel and Big Data, and SQL fills that gap nicely. Unfortunately, badly optimised MySQL databases have made people believe SQL can’t handle “big data” sets.
If you think you might have big data then you should have a read of my previous post, Why Big Data Needn’t feel so big.
Testing hypotheses with the data you have
You probably don’t have big data so I would advise you to start with an idea. Use your imagination and back the idea up with enough data to give it credence. Once you’ve implemented the idea, data can help prove its success (or not), and then the more data you gather throughout your marketing campaign, the more information you’ll have to guide your future marketing decisions.
In Scott Brinker’s article he argues that “The sensible answer for most companies is a balance of data analytics and human judgement.”