Todd shares how important it is to have good information underlying your reports and analyses.
Just a few years after I had launched my firm, I was designing and developing a system integration application for an apparel company. Even though this was a fashion apparel company, it had a decidedly old-school chief financial officer (CFO) – and I say that with the utmost respect. A former Marine, this CFO looked it – short hair, impeccable grooming, creased pants, pressed shirts, and highly polished shoes. In other words, buttoned down – and he ran his department the same way.
One day, a simple sign appeared on his door ...
I asked one of his managers what this meant. "It's Joe's new initiative. He wants us to make sure every analysis and report we prepare is TAC – Timely, Accurate, and Complete."
I thought this was interesting, filed the thought away, and moved on to the day's "To Do" list.
About 15 years later, TAC came to mind as we were discussing a potential data warehouse project with one of our clients. This client's industry had a number of standard metrics that were difficult to compute correctly. Internally, different groups had their own calculations of some of these metrics and – not surprisingly – the values did not always agree with each other. The reason? While each group used a similar calculation methodology, these methodologies were not exactly the same. Furthermore, the data used was obtained from different reports within the main business system. Sometimes, a report would be written for a particular purpose or user, and then another user would find it and use the data, without understanding the report's purpose. Consistency of information between reports was an issue.
These examples came to mind last week, when we obtained a new client. One of their major customers had just brought in another supplier, and we were told this was because our client could not provide customer-requested information on project status quickly enough for the customer's needs. Previously, improving the Timeliness of information did not appear Cost-Beneficial – now, with a competitor feeding from the same trough, the cost of lost business tipped the scales.
I began to think about how to measure information quality. A few Internet searches showed there was no generally accepted definition or measurement of "information quality." I did find several articles describing various aspects of information quality, so I have distilled these into the following characteristics:
Timely – Information needs to be available quickly enough to alert managers to issues, so they can take action while the issue is still small.
Accurate – Users presume a system will have various validity checks minimizing the risk of inaccurate data entry, and will have been tested to ensure calculations are accurate. What happens if two reports purportedly presenting the same information in different formats yield different results? Some authoritative source must exist to referee and resolve the differences.
Complete – All the information needed to make a particular decision has been presented, and nothing has been "cherry picked."
Consistent – Information is internally consistent within a single system (for example, a "Gender" field would contain only "M/F" values, and not a mix of "M/F" and "0/1" values), and information is consistent across multiple systems (for example, "Sales" in QuickBooks should agree with the total of "Sales by Customer" in a Customer Relationship Management system).
Cost-Beneficial – There is a cost to collecting, processing, presenting, maintaining, and storing information – thus, the cost of all this should be weighed against its value. Remember, information exists to help managers make decisions and take action – so, the bigger the impact of a decision or action, a larger investment to provide needed information can be justified.
Availability – Even if information possesses all the previous qualities, it is of no use sitting in an individual's private folder on a shared drive – it needs to be readily available to folks having a need for this information.
Ease of Use – Information needs to be easily accessed, reviewed, and analyzed by users.
Presentation Quality – The presentation of information needs to be appropriate to the user and its intended purpose. For example, a chart of "Sales by Region" on a dashboard is likely sufficient for an executive team's use, while more detail is likely needed by sales managers and representatives working to hit sales targets.
Relevant – Information needs to be pertinent to the decision at hand.
Trusted – Users need to know they can rely on the information available to them. An occasional small error should not jeopardize trust in the information. A major error or frequent small errors call the trustworthiness of the information – and the processes, technologies, and people behind this – into question.
You've undoubtedly experienced problems with one or more data quality characteristics. None come to mind? Think about the last time you did an Internet search – how would you assess the quality of the results?
High quality data has several benefits, including:
Better Business Decisions – Knowing which products consume more materials than specified by standards allows for corrective action.
Improved Regulatory Compliance – A $1 million error in a small subsidiary of a large corporation may not materially affect the overall consolidated results, yet the error does affect the results of that subsidiary and would cause an issue during an IRS audit.
Added Value to Customers – One way companies can differentiate themselves from competitors is to routinely deliver value-added information about the status of a production order or a project.
In looking at our current client projects, we are helping address information quality issues in all of them. Here are a few examples:
For a client switching from QuickBooks to an industry-specific ERP (Enterprise Resource Planning) package, the ERP system's import routines are turning up issues with the Completeness and Consistency of information coming out of QuickBooks.
For a client addressing a large inventory shrink, accountants are using data from a key system to make adjusting journal entries affecting inventory. While data from the key system has been tested for accuracy of its methodology and calculations, the manually-created journal entries have never been tested in the same fashion. The adjusting journal entries likely have an Accuracy issue, and thus are not achieving their intended result.
For a client facing a new competitor at a large customer, issues with the Timeliness, Availability, and Ease-of-Use of information on customer projects allowed a competitor to come to the table.
For a client overhauling its IS (Information Systems) infrastructure, Inconsistent classification of inventory items means reports and analyses are not always accurate.
What's the quality of the information you receive? How well does it help you make better business decisions? How might the quality of information be improved, so you can achieve more benefits from it?
Need help answering questions like these? Please call or email anytime.
Todd L. Herman
Would someone you know find this article useful?