Self-Service Data Prep Tools vs Enterprise-Level Solutions? 6 Lessons Learned
A detailed comparison between self-service data preparation tools and enterprise-level solutions, covering business strategy, accessible tools and solutions and more.
There are volumes to say about the often challenged relationship between IT and “the business” that has existed since IT became IT. Centralization, decentralization, self-service tools and applications, enterprise tools and applications – the pendulum swings again and again.
You’d think by now that we’d get it. There is no one all-encompassing data management or BI solution that will satisfy all of your data related requirements. Whether driven by enterprise requirements for security, process and control -- or by business needs for flexibility, agility or ease of access -- as soon as solutions are implemented, they are challenged to remain current with consistently evolving business needs. IT usually finds itself with the difficult task of supporting multiple sets of business and enterprise needs in an endless cycle.
Senior management teams need to not only accept, but embrace, the balance between competing sets of corporate needs for governance, risk mitigation, and management of all technical and data assets; and end-users’ needs for agility, flexibility, timeliness and accessibility to those corporate data assets.
One side of the pendulum swing is bringing us exciting, powerful and cutting-edge technologies in enterprise data management, risk management, and different types of BI, analytics and reporting systems. Considerable amounts of money, time and other resources are being spent on these, but in many cases we can still ask the questions: Are they delivering ROI? Are they agile, flexible, accessible and delivering benefit to the business, financial and data analysts who run the day-to-day business?
The decentralized side of the pendulum swing is delivering strengths and benefits in the growth of self-service data tools, visualization tools and more. But this is also not without its share of issues. In a world of data (in)security, one might ask: why does almost every company on the planet still have from a few to armies of business users using unauditable, error-prone and unprotected spreadsheets as the currency for file/data exchange, data manipulation, and reporting?
Why is this not headline news? Certainly, no enterprise wants their data vulnerabilities exposed. So, I understand why senior management and IT don’t want to publicize how large and potentially impactful this not-so-secret secret is. But with risk mitigation becoming more and more critical in every enterprise, especially publicly traded companies, this issue will not go away.
So, what now?
6 lessons learned on the topic of data tools and solutions:
- No matter how solid your enterprise data systems and solutions might be, current realities still show that data analysis and reporting often requires considerable data preparation (blending, cleansing, wrangling, transformation, normalization, etc.) before accurate analysis and reporting can take place. Don’t underestimate the value of robust and accessible data preparation tools to provide exponential time-savings and increased data quality.
- Balance strategic and tactical: With so many different types and sources of data, different types of users, diverse and consistently changing business needs, a successful “one size fits all” enterprise data solution is unlikely. Some data topics and issues should be addressed at the strategic / enterprise level, while others are more specific to individual business units – different people, different time frames, different priorities, risks and/or costs. Find balance in your organization for both strategic and tactical solutions to address your data issues.
- Don’t be afraid to start small and build from successes. There is no way you’ll be able to address everything at once. So don’t let the magnitude of the larger tasks stop you from addressing one key priority at a time. A successfuland scalable combination of tools and expertise can be replicated across the organization much more easily and with a lot less risk (and cost) than an extensive implementation.
- Business users need to have tools and solutions that are accessible (easy for them to learn and use). If corporate or IT-sanctioned tools are not accessible to the business users, then the business users will be very creative at finding alternatives. That said, IT should not be expected to learn and support every tool or solution proposed by the business, but neither should corporate policy limit the business from identifying appropriate solutions to its problems, so long as those solutions adhere to corporate data integrity and/or governance standards.
- Corporate data governance standards and criteria in the selection, implementation, and use of data tools are as important, if not more so, as the choice of tool itself. Create them, and ensure that all tools used adhere to consistent corporate data governance standards. Your criteria will be those that are most important to your organization. However, don’t underestimate the importance of: (a) Ease of implementation, learning and use of tools that are appropriate for the proposed users and their tasks at hand, and (b) Full transparency and visibility into all data manipulation processes, data relationships, and data transformations executed within the tool.
- As ubiquitous as they are, spreadsheets for some data work don’t appear to be going away. But there are numerous tools on the market that were developed to address areas where spreadsheets fall short. Such tools include: (a) spreadsheet management for the files themselves; (b) managing or tracking the relationships of data within spreadsheets; and, (c) those for simplifying and automating the time-consuming and error-prone data cleansing, preparation and reporting tasks that people default to spreadsheets for in the first place. Find them and use them. This is an area where you can find immediate and calculable ROI.
Bio: David Lefkowich is the Chief Marketing and Business Development Officer for FreeSight Software, a data integration, preparation, cleansing, query and reporting tool.
- Data Retrieval and Cleaning: Tracking Migratory Patterns
- Improve Data Science Productivity with Anaconda Enterprise
- Text Mining on the Command Line