4 things to consider when developing your data strategy

Organizations often struggle when coming up with their data strategy. In the end we need to be practical about what we can actually do and any effective data strategy has to keep focus on maximizing return on investment.

Here are 4 specific things to consider

1. Organizational Goals

2. Data Sophistication

3. Budget/Resources

4. Readiness of Your Audience

1. Organizational Goals

First let's look at the definition of strategy

strat·e·gy

/ˈstradəjē/

A plan of action or policy designed to achieve a major or overall aim.

This may seem like an obvious one, but in order to plan on how to get somewhere you need to know where you are going.

There is an infinite amount of data in your business and in order to decide what to focus on, start with the destination.

What are your organization's goals?
What business processes do you need to have in place to achieve these organizational goals?
What analytics/metrics do these business processes need in order to achieve these organizational goals?
What data do these analytics/metrics need to support these business processes in achieving these business goals?

Examine your current data gathering and reporting mechanisms and confirm: do we have all the data we need to feed to our analytics to support our business processes in achieving our goals?

And going the other direction, if you cannot clearly show how a particular set of data you are collecting results in achieving your organizational goals, there's a pretty good chance that either your goals need to be refined or that data needs to be deprioritized.

Each level should fully feed the next, no more and no less.

2. Data Sophistication Big Data, Machine Learning, Predictive Analytics: it's all quite heady stuff. Data science has enormous potential to transform your business.

But you need to crawl before you can walk: the theoretical promise of data science may not jive with the messy reality of where you actually are with your data.

For example, It's all fine and dandy to talk about predictive analytics but if you are tracking core business data with excel, you may be better off working on your data entry/capture mechanisms.

Or perhaps you are struggling with poor data quality and the metrics you report do not accurately reflect reality. In that case a first course of action may be to spend time and resources in improving your organizations data culture.

Take a beat and critically examine your data flow from data entry to reporting. Garbage in is garbage out so don't spend time building a perfect analytics machine when your data is not up to snuff.

3. Budget/Resources

This is another obvious one. We live in a finite world with finite resources. The best theoretical technical solution is pointless if you cannot afford to actually implement it. You need to deploy your finances and resources effectively, again with a view towards the end goal.

Sometimes that means taking one strand of your business and building out your data strategy from data capture to actionable analytics, putting together all the components from soup to nuts for just that one strand.

This is a good approach to take when you need a proof of concept to sell the idea to other business units.

On the other hand, with savvier or more sophisticated data users it may be better to build out a core infrastructure that focuses on improving data quality more broadly or on making data more easily available to your end users.

In that case you need to make sure your end users have their expectations set so that they know what the quality of your data allows and doesn't allow.

For example, if your data quality is poor but in a consistent way then perhaps you can still do trend analysis even if you can't measure absolute state.

4. Readiness of your audience When you are making changes or putting in place something new, there is inevitably going to be resistance. Some people are going to be more open than others.

It's human nature.

People often view change from the standpoint of wanting to stick with the devil they know.

You need to push people but if you stray too far from the familiar you risk tarring your new data initiative with the fear of change. You will have to judge the openness of your staff to change.

Certain people are going to be easier to convince than others. You are going to have your hands full doing the work so make your life easier by focusing on the groups or teams that are more open to your idea.

Perhaps it's because you have a more friendly relationship with those individuals, perhaps they are a team that normally gets short shrift and grateful for some attention.

Whatever the reason, starting with a team that is more open to things will help smooth the way. You will be burning social capital as you figure things out so start with those folks with whom you have capital to burn.

One additional caveat is that it is best if any pilot initiative is not at all customer facing. Staff can deal with hiccups in the sausage making process but customers are less forgiving, they just want things to work.

Final Thoughts In the end, having an effective data strategy is like having an effective anything strategy: you start with a vision and then anchor the implementation in reality.

The balance between the two is where we find success.