Data Strategies

3 More Secrets of Successful Data Strategies

A few weeks ago we talked about the top 3 factors of a successful data strategy.  
I’ve had a number of people ask me for more—it’s a broad topic that’s hard to summarize in three bullet points. These people work in different industries and lead companies from 200 employees to 200,000. While the advice is general it’s also broadly applicable.
Since we’re still interested in the success of our data programs, and since there seems to be an appetite for more ideas, I thought I’d expand on the list and share three more ways to help your data strategy be successful.
In the last post, we talked about treating data as a utility, bringing the data to the people, and driving the program from the business. I’ve seen these approaches be successful time and time again. But there are a number of other factors for success—so let’s take a look at some of them.
Since these are broad ideas, I encourage you to think about how they apply to your specific business. If you have other ideas, please share them in the comments!
We ended the first article in this series by talking about finding value. So, how do we find that valuable problem? In general, there are two ways. 

Do something better

The easiest way is to identify and implement something that can be done better. You may know where the opportunities to improve are already. If not, that’s fine—your business users certainly do.
The best place to start is to look at their workflow. Talk to the SVP of Sales, the CFO, and the CMO. Look at how their people spend their time, and where they feel it’s wasted. Some questions to ask: Where’s the frustration? What takes them longer to do than it should? Where is the time wasted? What causes customers the most concern?

Some examples

Although these types of “do it better” opportunities are often obvious, here are a few I’ve run into repeatedly over time.
Sales. In any industry that provides an ongoing service to customers, salespeople will often meet with them to provide information about historical performance. How often do those customers ask a question that the salesperson can’t answer with an existing report? What’s the turnaround time to answer that question? If the answer is anything less than one minute, you’ve found a problem to solve.
Marketing. How successful is your current campaign? This is often answered by a report that was defined well in advance of the campaign. Did you think of an attribute that you’d like to “slice and dice” by after the fact? Sounds like a new report request.  
How many days will that take? One customer I worked with years ago realized than an in-flight campaign was actually causing them to convert significantly fewer customers on their website, and stopped that campaign immediately. The estimated value of knowing the answer now? Over $1,000,000.
Finance. Much of the analysis done in finance is static and repeating. There are only so many ways to represent a balance sheet, after all. But many of your finance folks focus on more ad-hoc analysis. How much time is spent finding expense report outliers? How do they find out why software expenses are so much higher this quarter? It’s not unusual for that type of follow-on financial question to take days or more—a great opportunity for workflow improvement.

Do something more

The most valuable ideas happen when we reinvent how things are done, not refine them. Apple is great at this. At the height of the iPod’s popularity, they effectively killed the category by introducing the iPhone. Nobody can say now that that wasn’t a brilliant move.
Apple is also famous for not doing user group studies. Apple believes that if you ask users what they want, they’ll only tell you how to improve what they have. Increment improvement is valuable, but won’t result in exponential improvement.
When it comes to data, doing something more often involves creating a new revenue stream. When I talk to CDOs about this, the most common feedback I hear is, “We can’t create a revenue stream with data, there are privacy issues and we’re barely using our data to understand our own business.”  
On the surface this makes sense, but in practice there are many ways to monetize data that don’t run afoul of privacy issues.
Resell aggregated data to your market. When data is rolled up to a high enough level, it’s entirely anonymous. For example, if you are a cable TV provider, resell viewing figures for shows, in near-realtime if possible.  
Resell aggregated data to another market. Many investors rely on regularly published Consumer Price Index numbers from the government for investing decisions. If you run a retail operation, you may have a lead on estimated figures a month or more in advance based on your rolled up sales numbers. There’s a market for that information.
Customer self-service. Do you provide information back to your individual customers about your product or service? How happy are they with that service? How many ad-hoc requests do you get that aren’t served by current reports? What’s the turnaround time for that, and the expense? Getting the data to your customers instead of making them go to the data—as we discussed before—is potentially an extremely valuable addition that many customers would pay a premium for.

Always ask for data

The ultimate goal of any data strategy is to change the culture of an organization by leveraging data to make more timely, valuable, and relevant decisions as quickly as possible. Changing the culture of a company takes time, but it really just consists of a series of small actions done consistently. The most important small action? Always ask for data.
When someone tells you they’re going to try something new, ask them what data they’ll use to measure the success. When someone tells you that a number they’ve provided looks good, ask them how they know. Have they put it in historical context so that you know it’s really good? When you’re told that a decision has been made to choose a specific product, approach, or path, ask to see the data behind the decision.
Ask often enough, and people will start to look at and prepare the data before they come to you. And that’s what you really want, isn’t it? 
Netflix is very good at this. They rarely make a decision about anything a customer sees without data to back it up. That cover art for the new show you’re dying to see? It’s the result of many different A/B tests to determine the best visual to show to get people just like you to watch a show. There’s nothing casual about it.

Have you tried any of these ideas?

How have you been successful? What else should your peers look out for?