The Data Strategy: Alternatives and Outcomes

Posted on by Christopher Berry

A Data Strategy is a set of choices that reinforce each other, that are difficult for competitors to replicate, and that generate a sustainable competitive advantage.

A Data Strategy is set of choices about data. The benefits are the short-run outcomes. A sustained competitive advantage is a medium to long-run outcome.

This post is about the alternatives and outcomes involved in a data strategy.

Alternatives

Alternatives can be really obvious. The most obvious ones involve fiddling with the settings on instruments that are available to you. In the face of falling intelligence, you can chose to increase the number of spreadsheets, or decrease the number of spreadsheets.

And that’s really obvious. Because data, in most organizations, lives in spreadsheets.

Less obvious alternatives involve selecting tools. Maybe spreadsheets aren’t the right tool to fiddle with. What if the delivery mechanism is something to fiddle with?

And then the even less obvious alternative involve stances, or ways of thinking. Maybe the market for data isn’t the problem. Maybe it’s the way that you’re thinking about the data market is?

Searching for Alternatives

Sometimes you just accept the alternatives that are obvious. And that’s fine.

People generally aren’t prone to seeking out more alternatives without a reason or some propellant. Going on a hunt, or a search, for alternatives takes some effort. Doing a serious search takes serious effort. In general, the amount of effort that goes into a search is proportional to the pain that the problem is causing. If everything is fine, in general, you won’t search.

You won’t visit Forrester for a vendor list. You won’t visit Gartner looking for the magic quadrant. You won’t talk to your friend in the startup community that knows a lot about new technology.

Satisfaction with the status quo causes stasis. There is very little reason to engage in a search if things are going well.

Searching for Alternatives in Data Strategy

Most alternatives begin and end with technology:

  • The spreadsheet.
  • The Adobe SiteCatalyst.
  • The Tableau server.

Several alternatives begin and end with a single person:

  • The intern that has some time to put together some spreadsheets.
  • The account manager.
  • The analyst.

It’s rare that process solutions are ever sought:

  • The double entry.
  • The audit trail.
  • The process of separating data models from controller logic from views.

Many searches end with simplistic frameworks:

  • Have sex with data!
  • The Nine V’s of Big Data!
  • The Balanced Scorecard!

And, it’s even rarer that processes ever meet technology ever meet people:

  • The Analytics Center of Excellence.
  • The Federated Model.
  • The Unitary Model.

So, if there’s a reason why so many people sound the same, it’s because they’re all ending their searches at different points, and that’s generally predictive of just how frustraged they are.

Searching for Outcomes

We don’t spend much time talking about outcomes. What’s the outcome of a data strategy? What is desired?

It’s very common to get a single word answers to those questions:

  • Proof.
  • Answers.
  • Insights.

And those are outcomes. One word answers are terrific communication vehicles in our attention-starved world.

It is far rarer to get answers like:

  • Higher profit.
  • Less churn.
  • Sustained competitive advantage.

These are harder answers. It’s harder because the number of hops in logic is longer. In general, an argument in the form If A then B is far more effective than an argument in the form If A then B then C.

Here’s an example:

Demand is abstracted out as a curve on a chart comparing price and quantity. That’s why it’s called a demand curve. If you can use information about consumers to charge different prices, based on where they are along that demand curve, you will get more profit.

Does the data strategy support that outcome? Where does data about price come into the mix?

What about about less churn? What if you could use data to predict which consumers were likely to defect to a competitor, or drop the technology altogether? What would you do to intercept that behavior and reduce it? Retention is also a curve.

And then the ultimate so-what. Many firms in competitive industry make use of price discrimination and loyalty/retention programs. What natural advantages does your firm have, with respect to data, have? What system of activities would be very difficult for any of those firms to replicate?

It’s sort of hard to get to any of the good outcomes with the simple alternatives, isn’t?

Data Strategy: Alternatives and Outcomes

Every firm faces constraint. There aren’t enough resources, even if there were unlimited monetary funds (we don’t all work for the Fed, and even then….), to do everything well. Data strategy is about selecting the appropriate alternatives in pursuit of an outcome.

There’s a good school of thought that’s built around providing situational awareness to people who are authorized, as a matter of policy, to make decisions based on a situation. The argument goes that if you ask these folk what data would make a real impact on making better decisions, they’ll tell you. Faster situational awareness is fairly difficult to scale by way of the existing spreadsheet regime or with a single person, so, there’s a different sequence of alternatives that are required over time.

There’s another school of thought that’s built around taking a big leap all at once. Think about the Japanese Meiji Restoration experience. Japan decided to remake itself in the mid-19th century. They took the best of German labor policy with the best of French administration policy with the best of American industrial policy and they jumped forward. Or, you look at how the sub-saharan information technology industry is leap-frogging India. That school is convinced that a firm can compete on analytics faster than other firms can, all in one big leap. Take a little bit of Google, a little bit of Amazon, and little bit of Netflix, and presto chango, you’ve done it.

Ultimately, the identification of an outcome, and working out the data strategy from that perspective, is far more likely to generate a better result than incremental reactionary policies. The alternatives involved in getting all the way there might be opaque. But, if it was obvious, everybody would know it. And there’d be little competitive advantage to be had.

It’s in the mix

There isn’t a one-size-fits-all approach that fits well, because different firms in different industries with different fluid situations exist. But you can pick something that does work for you.