Have you heard about companies with a "Data-Driven Culture" – what does this mean, and how do you become one?
A while back, we were retained by a company that knew their accounting system was broken, and suspected the accounting data being produced was inaccurate. Our initial assessment of the accounting and finance areas confirmed their suspicions. We then made numerous detailed recommendations to address the identified issues, including selecting and implementing a new accounting system.
The client was hopeful. Our assessment mapped out a path to produce dashboards of timely and accurate information to help track and evaluate the many new products and services they would be introducing.
At the time, I didn't know a good, concise term for what this company wanted to achieve. Recently, one of my associates shared an article with just the right term – our client wanted to build a "Data-Driven Culture."
Foundation for a Data-Driven Culture
The article, The Four Key Pillars to Fostering a Data-Driven Culture, discusses four essential components identified by the author, Brent Dykes. Here is my summary of his pillars:
- Shift the MINDSET – Both diligence and patience are required to "steer your team in a new direction."
- Strengthen the SKILLSET – Your existing staff already possess valuable industry experience and domain knowledge. Help them build their skills in data literacy and data storytelling.
- Sharpen the TOOLSET – Automated data pulls, data cleaning, and data integration are key tools and a prerequisite for data analysis. Pruning the variety of systems and data analysis tools provides a common data analysis platform. This common platform enables you to develop a "single version of the truth" – a clean set of financial and operating metrics everyone knows and trusts. From this base, you can confidently leverage specific dashboards for various purposes, such as using a dashboard instead of a presentation deck.
- Solidify the DATASET – Once the company's key priorities are clearly defined and communicated, you can collect and analyze data aligned with these goals. At this point, your people will view the data as a key company asset.
Fortunately, our client already has several attributes of these four key pillars.
- The CEO has consistently supported efforts to implement our recommendations, as she works to improve her team’s effectiveness.
- Top executives and many top managers have both domain knowledge and data literacy.
- All personnel want to have understandable and actionable information.
What they lacked, however, was a solid foundation of underlying data and processes for the four pillars. The data must be clean – timely, accurate, complete, and relevant. The processes – whether performed by people or by a system – used to produce the data must be both repeatable and performed accurately.
Poor Systems + Convoluted Processes = Data Nightmare
For the past several months, I've been working with this client to implement a new accounting system. Because the old system didn't have a good financial statement report writer, information was imported into a separate report-writing product to manipulate and dress up the financial data. Only one person in the company knew how to use this report-writing product – a risk area itself.
Even with the separate product, the resulting financial statements were neither specific enough nor timely enough to help executives run the business. When it came time to convert the data, we had to decide what was the “real” information – the information in the old accounting system, or that in the report-writing product. The answer was the latter, which made loading the data into the new system very challenging.
Cleaning Data Is an Iterative Process
Never underestimate the effort required to clean and load existing data into a new system. In my experience, this has NEVER been a "one and done" effort – rather, it has required numerous steps, each repeated several times. Even if the existing data is relatively clean, the steps to clean and load data are still repeated several times prior to "Go Live" because you only get one chance to load the data and go live "for real."
In cases like my client’s, when the existing data is known or suspected to be dirty, there is only so much cleanup that can be done by a data conversion team. Assumptions had to be made about what information was likely to be accurate or not, and these assumptions were communicated broadly so everyone knew the cleaned-up data would be better, yet not totally accurate.
Ultimately, as with this client, turning dirty data into "acceptable" or "good enough" data that is “directionally correct" can serve as an acceptable foundation for the four pillars.
Once the data was cleaned as much as possible and loaded into the new system, we were able to produce reports showing the executives – for the first time ever – revenues and expenses by their major products, services, and cost centers. While the executives were thrilled to have these new reports and additional information, they also found several examples of things the accounting staff inadvertently had been doing wrong for years.
Developing Processes for Accuracy and Repeatability
The staff never realized they were making these mistakes because the prior convoluted reporting process had hidden them. Broadly, all the mistakes had to do with using the wrong revenue account for a particular product or service.
To address this, we worked to help error-proof the accounting processes by creating sets of rules specifying valid revenue accounts for the major types of products and services. After reviewing the new rules with the executive team, they found even more ways to simplify and strengthen them. We refined the rules, and reviewed them with the top two company executives, who green-lighted building dashboards in the new system their accounting staff could use to identify – and correct – transactions not conforming to these rules.
Almost There!
My client has cleaned the data for the fiscal year being audited, and staff are now cleaning the data for the current fiscal year. While cleaning the prior and current years' accounting data has been difficult and time-consuming, it has been essential to ensure information is comparable between years. Furthermore, putting in place tools to help detect – and even better, prevent – similar errors demonstrates that the processes creating dirty data are being fixed.
While the executives know they are still not where they want to be, they are seeing the improvements needed to put them on the path to achieving a solid foundation for the four pillars of a Data-Driven Culture.
Sincerely,
Todd L. Herman
