The demand for data preparation solutions is at an all-time high, and it's primarily driven by the demand for self-service analytics. Ten years ago, if you were a business leader that wanted to get more in-depth information on a particular KPI, you would typically issue a reporting request to IT and wait some indeterminate time for a dashboard or report to be constructed – which hopefully matched your needs.
Things are very different today.
Cost, complexity swapped for self-service
The cost and complexity of enabling analytics have dropped, but the bigger change is the shift in capability. The business is no longer waiting for IT to provision analytics requests. Instead, business users are increasingly acquiring the ability to go down the self-service route and do the processing themselves.
Much has been written about the dangers of self-service computing and the risk of providing unfettered access to potentially sensitive data sources. These are valid concerns, and the CDO/CIO partnership must work to ensure that the right balance between governance and innovation is preserved.
But in the midst of all these discussions about data privacy, stewardship and protection, I think we sometimes miss an important point: The deeper understanding the business community has about the quality of its data, the more inclined they will be to manage and improve it.
For far too long, data quality management has been left to the IT team or a central data quality team or stewardship group. Sure, the business is often involved, but often at an arm's distance.
Self-service can lead to better data quality, governance
With self-service data preparation and do-it-yourself (DIY) analytics, the company will gain a much clearer view of where data quality affects the insights they are so desperate to gather. If business users are simply spoon-fed analytics reports, they often fail to appreciate the quality of the underlying data. By doing the processing themselves, they will quickly see how the different rules and levels of data quality influence the outcome of high-quality decision making.
In a previous role, we undertook a similar self-service tactic in an automotive market research firm. We trained the business users in the finer arts of data processing and manipulation to remove the IT bottleneck that was causing excessive lead times for custom reporting requests from clients.
Here's what happened as a result of this self-service experiment. The users, many of whom only had basic Excel skills before the classes, soon began to understand the impact of the codes and values they entered in the data. They started to coach new hires, and they implemented improved policies and training so that the reports they needed at the end of the information chain would not be subject to gaps, duplicates and invalid entries.
In my experience, self-service analytics can be a driver for greater data quality and data governance – not an inhibitor.
What are your experiences? Are they the unchallenged Wild West of information management, or a force for good?