Eventually the day comes when an existing data mart or warehouse needs to be re-examined. Maybe new workloads are desired, new business models arise, a new platform is considered, or new supporting tools in the environment are wanted. Especially when a new platform is part of the equation, many in charge think first of “lift and shift” as the approach. Taking as much existing data and processes as possible and keeping them virtually unchanged as things are moved onto new foundations is almost always the cost-saving approach—for the immediate short term. There may be good reasons to perform a “lift and shift” as a first step. However, if efforts stop as soon as this first step is completed, the organization has likely chosen not to leverage many new opportunities. New platforms operate differently than old platforms. Functions and tasks that are performed well, or poorly, change between them. If one is using unchanged algorithms, and unchanged data structures, then chances are high one is not truly improving circumstances as much as they might.
Approaching things from a fresh perspective is a good thing when new platforms are involved. Each platform has different strengths and weaknesses. What was done to optimize performance on the old platform may not be necessary on the new platform, or the new platform may have that extra something allowing for leveraging more flexibility than was done previously. As an example, aggregations are generally done to improve query performance; a modern data platform may be able to provide great performance without needing a physical aggregation, or alternately an aggregate, if done in a slightly different fashion, may provide even more flexibility for users than they had previously. Ideally, when considering a new platform some proofs of concept tasks were executed that played around with the largest pain points to be addressed by the change. In this way, the proof of concept would have alerted the team as to what areas might be the most important to rework and give some direction as to how that improvement will be accomplished.
New data platforms and supporting tools will often require a re-tooling of the organization’s staff. Certainly, if all of one’s staff are consultants, one may swap one group of skills for another, but it now may take the new consultants time to learn about the business. If possible, taking the time to both educate your team and allow them the time for some hands-on efforts to become more intimate with the new tools is an organic and wholistic approach that helps everyone. Effectively, an organization could perform a long-term proof of concept, start using and building up things utilizing the new tools/platform with the expectation that once the team has learned how to use the new tools well, the proof-of-concept items will be re-done more formally. If there are pieces of the proof of concept that do not need re-work, then great, one has a head start.
“Lift and shift” is not a bad starting approach, but if possible an organization should be more surgical. New tools and new platforms are generally brought in to improve overall performance. Picking the most critical pain areas and reworking them right away as part of the conversion efforts would be good. But even if such reworking is phase two, that may be fine too. Just do not set up circumstances where a “lift and shift” is done and then ignored. Without taking advantage of the beneficial qualities of a new tool, the organization is losing out on improvements offered by the changes they are making. It is like dressing up for the ball, arriving, and then just standing by the wall and never dancing.