19 Mar Is it time to go on a data diet?
by Kevin Willemse
Since it’s not the first week of January, chances are your New Year’s resolution to exercise more and follow healthier eating habits is nothing more than a long-forgotten memory. But that’s okay, because we’ve got a diet plan that will improve strength, supercharge speed, and maximize vitality. With just a few simple steps, we can eradicate those pesky pounds holding you back and get you back into peak performance mode.*
Well, your applications stack, that is. We call it the data diet.
This isn’t a blanket statement suggesting you should indiscriminately shed 10% of your company’s data weight overnight. Not only would that be catastrophic, it’s also probably illegal. But we are saying that in order to keep your processes and systems lean and in good health—guaranteeing a longer lifespan and improved capability—you should consider the same approach you might adopt for your own personal fitness. Here are Simplus’ six steps to get your data diet started:
*Individual results may vary. Always consult a professional. The data diet may not be for everyone.
Have a goal
It’s quite pointless to spend time and effort on any endeavor without establishing a clear intent and outcome. How will you know what success looks and feels like? Or how, when, and where to measure it? Spend some time strategizing why you are trimming back your data hoard, and this will result in or promote desirable outcomes. Your goals could be more efficient data capturing by end users, simpler integration between systems, improved reporting capabilities, faster processing times, or other benefits delivered by a leaner data waistline.
Consult a professional
As much as we think we know what’s best for our health, an expert nutritionist or simple doctor’s consultation can prescribe the best (and safest!) approach to reach our goals, as well as describe realistic expectations. When dealing with your organization’s data, unless you are intimately familiar with and employ internal resources to expertly maintain every system, integration layer, end user experience, BI tool, reporting platform, and other consumer point of data (a.k.a. “everything”), you should consider enlisting the services of a professional systems, data, and/or application services organization to help you navigate a path and form a plan that produces desirable results.
We may feel we don’t need help and resist it, but these specialists know exactly where to look for areas of potential bloat and waste and also how to eliminate them effectively without causing undue harm to their host—the entire organization.
Create a plan
Now that you’ve established the “why and the how,” look to the “when and the what.” Like an eating or exercise regime, set up tasks and timelines to address areas that you have targeted for reduction. Rather than attacking your entire database with a broad approach, zero in on attainable short-term goals, such as specific tables, objects, repositories or areas of your warehouse. Consider grouping these by functional user group or system.
Not only does this provide a feeling of satisfaction as you progress incrementally, it allows you to perform targeted tests upon completion of each task to ensure no dependencies have been broken between systems or user functionality has been lost. It is then far easier to identify and rectify these than having to undo the damage caused by a large or intertwined shedding of data obesity.
Find a workout partner
Simply put, don’t do the data diet alone. Identify all those users potentially impacted by the proposed trim-down and (even if you think you know their systems and data needs thoroughly) enlist their help in identification and testing of the before-and-after results. You’re doing it for them as well as for you, so involve them in your efforts and share your commitment to achieving the best results.
Be ready to change your plan
If a prescribed diet leaves you in worse shape or has adverse side effects, you should stop (and consult your physician) immediately to avoid further damage. Fortunately, a properly executed data diet means being able to undo any harm quickly and without any lasting impact; this means having regular, secure, tested backups and rollback plans in place before even thinking about stepping on the scale. Furthermore, even if these procedures are in place, perform rigorous disaster recovery exercises on datasets to ensure they are performing as assumed—and if you don’t have these exercises, get them ready first! Remember that, unlike the human body, data is not a self-repairing machine.
Change your habits
“The plan works if you work it,” the old saying goes. The same is true for data governance and maintenance. It doesn’t help spending months of effort transforming into the ideal “you” only to revert back to junk food and couch binges the moment you hit your target. The data slimming regime includes identifying and removing the causes of bloat in the first place. These are usually identified as redundant data fields, suboptimal processes, poorly enforced user rights, or piles of legacy data from deprecated integrations, rogue projects, or replication. Once you have completed the exercises of identifying and removing unwanted data symptoms (securely), cut out their cause straight away. It doesn’t help to simply keep the bad stuff out your body; get it out of the fridge as well.
Kevin is a Director in Simplus’ Strategic Advisory Practice, focused on bringing valuable transformation to customers looking to maximize their investment in Salesforce.com alongside their integrated systems and business processes.