Let’s assume you have managed to build a consensus that data-driven business requires previous planning of which data will be driving.
So far, from a non-IT point of view, you are just a Creator of Problems, spoiling your company’s chances to partake in business’ Next Big Thing. Business as usual.
Brace yourself, because you’ll have to rub it in and question:
how are we going to collect those data anyway?
CIOs need to be less about technology and more about what uses of technology can improve the business. IF it’s not too late.
Make no mistake, the answer to this question is not a requirement to be provided by an internal client, be it marketing or anyone else. The answer to this question is but one of the very reasons why a company needs in-house IT, and can only come from IT, or at least with a heavy involvement from IT. Internal clients have no need to understand information technology beyond the ability to use it for their own purposes, any more than IT needs to understand corporate law when setting up a shared document folder for Legal or HR.
Only a CIO has (or should have) a bird’s eye view of a company’s processes, because IT is the only hub of all company’s data. It’s about time CIOs stop acting as a mere (outsourceable) technology-driven repository and starts thinking and acting like the Chief Data Officer they need to become: someone who understands, promotes, extracts and safeguards the business value of data. And who, as a result, can provide advice on how and where to harvest the data the company needs.
[Note: as it happens, IT has been overtaken by events and Big Data is being led by function owners, i.e. by anyone except those who should have the skills and perspective to drive it and the impartiality to produce neutral data. We’ll likely see lots of Big Data and little data-driven business. So goodbye, IT, and good riddance. We’ll talk about this another time.]
You do not get to do data-driven business without knowing where data will come from and a deep understanding of your business processes. Data can come from any function, from HR to Sales. Even transit at the coffee machine is data you can work on. Data can tell you a lot, depending on what you have, for instance regarding sales:
- got order data? You can extract best-selling time of day, or day of week, best product combinations (A sells more with B than with C) and other hidden seasonalities. Also, do you know how long a client takes to place an order? Could you make it faster?
- product preference data. You know that product A outsells B. But if your margin comes from B, you either change your price structure, or investigate why A sells better. You need better salesperson training? Different product placement in the shop? Better marketing of A’s qualities?
- order-to-dispatch times. See how reactive your warehouse is
- dispatch-to-delivery times. See how your logistics can improve
- customer information? Build “classes” (small spenders, big spenders, split by age, sex, region, or whatever you see fit) of customers and apply different sales conditions to maximise revenue; devise incentives to increase sales (give 1 product free for every X bought to customers who abitually buy X-n for small n)
- shop access information? Work on product placement, lighting, opening times, offer times (maybe specials score better at off-peak hours?)
- salesperson performance? Forget raw sales data: see who sales what best when (and possibly to whom), and build on that. Maybe Tim is a bad overall seller, but outsells all others with specific customer segments.
Can Big Data help you with non-sales-related intellectual work? Yes it can. Here, too, formalised processes will pay off nicely:
- procedure completion times: slice by employee, by procedure time, by time of day, by workload
- management overhead: calculate the monetary cost of all meetings, internal memos and reporting
- actual productivity: white collar productivity is difficult to gauge, but you can monitor how much time is spent on social networks, and devise a policy; you should use this information to give people the slack they need, not to take it away (of course, those who prefer spending the day on Facebook should either be in Marketing, in Client Services or in a different company)
- office planning: people who work together need to interact, but how do they do it? They page each other? use Skype across adjacent offices? Have to engage in office tourism just to synchronise efforts? If you can track movements across the building, you can answer to these questions.
These are just examples. The actual data you are already collecting can suggest all these investigations and more. Also, working on data you have will prompt you to devise ways to collect other data. It’s a self-feeding process. You decide what your business needs to know about itself.
But, you have to decide beforehand: collecting data without knowledge of their intended use produces data of poor quality: input errors, missing fields, fields containing boilerplate values.
“But, what about the data I already have?”
As a general rule, you should consider them as good as scrap. Data quality does not come for free. Take a look at your CRM. I mean, a real good look. Any missing fields? Any boilerplate values? In general, CRMs are a data nightmare, especially if you employ sales agents who traditionally consider customers their own property.
Now, cross those data with Sales data: chances are they will not be aligned. This is what I mean by “decide where data will come from”. Data-driven knowledge does not come for free.
You need to make the company aware it will be collecting specific data for specific purposes.
- get function owners to identify the data sources they have and what they contain
- find out what you want to know that these data may tell you
- get IT to report on the data quality of existing data
- find out what you want to know that these data may not
- identify new data sources for point 3
- devise a procedure for the correct collection of data
- task IT to clean up the data and set up controls to guarantee the required data quality from now on
- instruct all interested personnel to the value of the data they collect and the importance of correct collection