Organizations that were born digital are built around their IT platform, and all their business processes are IT-driven and data-powered. Every action, every decision, is based on the processing of data sets about users and customers, about usage patterns, external conditions, etc.
But not every organization was born digital. If you run a traditional organization, with various degrees of sophistication in its IT, how can you still transform this business into a digitalized business? How can you create new business opportunities, based on your digital assets?
Here are 5 steps you should take that digital leaders use.
Drive your business intelligence toward real-timeliness
Chances are, you are already collecting transactional data in a data warehouse or data mart, and you analyze it somehow. Maybe it's to decide which sales rep to promote of fire at the end of each quarter. Or it's to establish a list of customers to send the new catalog to. The first step toward a digital business is to obtain this insight on a real-time basis. Now, "real time" does not always mean sub-second! In fact, I have argued before that "right time" is a better term: get insight when this insight is relevant to impact the business.
Instead of just using your sales reps performance ranking for HR reasons, use them to send your best reps the best leads, and to provide a gamified challenge to these who usually do great but are having a (hopefully temporary) hard time. And use customer segmentation for in-app promotion and text-message-coupons.
Of course, all of this requires insight to be obtained in right time. Agility and instantaneity constitute one of the foundations of digitalization.
Inject analytics into all business processes
Obtaining insight from real-time analytics is only the beginning. Business processes need to be retooled to accept this insight as an input. When business intelligence was all the craze, this used to be referred to as operational business intelligence -- the concept has not changed, only the technology. Real-time analytics, rules engines, intelligent business process management tools, push-style communication protocols make it possible to inject and modify rules, to trigger and modify actions based on insight.
Always look for more data
Storage is cheap, and the newest breed of data platforms makes it really easy to amass data even if the purpose is not clear. You may not be ready to move your transactional systems, or your data warehousing infrastructure, away from the tried and proven mainframes or RDBMS they are running on. That's fine. But consider complementing them with a data lake based on Hadoop, and to dump into this lake data you would not have considered worth keeping in the past: access logs, GPS records, abandoned carts, call data records, customer complaint messages, etc.
All of this so-called "dark data" can be used for new insight, for new actions. You may not know which ones yet -- but you won't know until you have had a chance to look.
Explore new ways to use your data
With reason, give unrestricted access to all data to your analysts. I say "with reason", because some industries have specific regulations that apply, and all industries need to be careful about data privacy. But, assuming you can trust them, and the proper governance exists that would curb any abuse, let your experts explore the data. They will find ideas -- some that will work and some that won't.
It's always difficult to measure the return on investment of innovation. But innovation happens when people are let to pursue ideas. Google got it right, with their "20 percent time" program.
You may not have official data scientists. But a good business analyst, equipped with modern tools for data preparation/exploration, can achieve amazing results.
Release often, test all the time and fail fast
This last point is probably the most critical one. Digital organizations are agile. They always test innovations in real-life conditions. Once a feature is complete, go live with it and measure its effectiveness. If it does not yield the expected results, be ready to pull back -- revert to the previous version, remove the option, try something else. And whenever possible, test different alternatives in parallel to see which one works best.
The key is not to always get it right -- nobody does. The key to success is to fail fast, and change course before it's too late.
This story, "5 steps for transforming your business using data" was originally published by InfoWorld.