A friend of mine used to be a big fan of computer games. Most of all, he liked playing immersive simulation games, and would sometimes play deep into the night, totally engrossed in building cities or balancing digital economies. I always thought it was a lot of boring preparation which never really led to any action.
Recounting his latest adventure, my red-eyed friend would explain that when a fire broke out in his digital city, it was not the right moment to think about building fire stations, or training firefighters. Although it may have seemed dull to a casual observer, this part of the gameplay was actually a critical task: to discover and work out how to keep his simulated world on the rails – as well as planning for what might go wrong.
You could say the same thing about Digital Transformation projects. Too many of them fail for one simple reason – and it’s one that is avoidable: businesses don’t have their arms around their data.
This is a problem that is not going away: the scale and complexity of the challenge is growing. Even before the widespread rollout of 5G networks to collect more data than ever from the network edge, data sets are vast and growing.
On one hand, the effusive analogies for the value of this data in today’s digital businesses claim it is “the most valuable resource in the world” or talk about the race to collect and analyze data as the “new gold rush”. But the hard truth is that enterprise data it is only valuable when it can be leveraged effectively to solve business problems or provide answers to crucial businesses challenges.
Unfortunately, many businesses have a poor understanding of what data they hold, and even less of a grasp of its potential value. And the inconvenient truth is that many enterprise digital transformation projects fail simply because businesses didn’t spend enough time learning how to use their weapons: to ensure the information they needed, to meet business requirements, could be identified and effectively leveraged.
The potential cost of failure is increasing to the extent that businesses now can’t afford to get it wrong. According to analyst firm Forrester; “2020 will be a wake-up year for many firms, as the total cost of getting data wrong will become apparent”.
Everyone wants to be a winner in the race to transform. A key element in our strategy to give digital transformation projects the greatest chance of success is the multi-phase data discovery process. This is designed to ensure that customers have all their data “ducks in a row” prior to launching a new transformation initiative.
The consultative process always starts with an agreement of the scope of the project, which leads naturally to creating a wish list of the data required. Skip this step at your peril.
The do’s and don’ts of data discovery
Game saved. Now it’s time for the next crucial step in any successful digital transformation initiative – to get a clear understanding of the data that your organization has at its disposal. While that sounds quite straightforward, it can in fact be astonishingly complex. Underestimating the scale, complexity or importance of the data discovery phase undermines the success of many digital transformation projects.
Data discovery involves understanding where to find the data required for each specific transformation objective. You’ll have to be imaginative to think of all the places in an organization where data resides. For example, do you check your website traffic statistics? Have you ever screened your backup data to look for trends in your data?
Leave no stone unturned. Check virtual environments and any public or private clouds your organization uses. Dive into the mails your service department received from customers, look at customer delivery data , evaluate the sensor data from machines you delivered, and look at how long customers needed to wait on the line before your company answered their call.
Also, be aware that the data you seek will be spread across applications and there will be multiple copies duplicated in a plethora of backups, with many versions and duplicates to untangle. To ensure that the location of relevant data is identified from the start, across all locations, you need assessment tools that crawl over all your files to dig out anything needed for a transformation project.
For projects in industrial companies, you will also require IoT data, for example performance data collected from manufacturing machines or processes. But you’ll need a power up before you can get to this data, which must be translated for it to be understandable and useable outside of the operational technology (OT) machine environment.
We’re using the Fujitsu INTELLIEDGE appliance and gateway system to bridge this IT/OT chasm. It’s rugged enough to sit in non-IT locations where it can render the OT data for widespread IT use and has enough input ports to collect native data from multiple sources.
Classification, classification, classification
Once we’ve completed this foundational discovery process, the next steps are to map available data against desired business outcomes and classify it. Classification is another complex undertaking and is generally something where our customers need specialist help. Given the scale and intricacy of enterprise data today, this should not be a manual process, and establishing and running any automation requires significant expertise.
We start by identifying which records are important, which are not, and what additional input is needed to deliver the insights required, all within the context of the overarching project objectives. It’s like one of those video games where you need to collect all the items – scattered across a map – before you can unlock the next level.
Descriptions of the required information are turned into discovery policies – these are used by intelligent tools to identify and classify the topology of the data – essentially creating a 3D overview of its source. The data is scanned, and metadata is collected.
To do this, we implement additional sophisticated software – assessment tools that collect the metadata for all relevant structured and unstructured data – identifying sources, ownership, descriptions and dependencies. Yes, this is even more classification.
These powerful tools mean even a heavily distributed, complex hybrid IT infrastructure can be explored automatically. We take metadata and create automatic data classifications (sorry) using filters, search words and tags. We also generate detailed reports that visualize where different types of desired data can be found.
As part of the Fujitsu consultancy and exploration process, we deploy the most appropriate tools to assist us with this task – for example the APTARE IT Analytics platform from our strategic partner Veritas. Once complete, all the relevant data types and kinds are essentially indexed, and ready to be put to work.
Go slower to go faster, or speed ahead and go around in circles
It’s tempting to skimp on these crucial discovery and classification steps. But trying to implement a digital transformation project without them is like looking for a needle in a haystack. We recommend starting out slower for a better final result.
Not only will effective data discovery and classification boost a project’s chance of success, but a solid understanding of your data, where it is located and its potential value, will also be helpful in meeting a growing number of regulatory requirements.
Fujitsu supports this process at every step of the way. Our one-stop-shop approach tailors each process to individual customers’ needs, combining extensive systems integration expertise with the most appropriate combination of market-leading tools from our strategic partners.
This powerful combination means that even those businesses who start out with no idea of how much data they have, or its potential value can establish the data foundations that allow them to strike “gold” creating new value from existing data. And that means you’ve got more chance of getting your name in the hall of fame, instead of having to insert coins to continue.