How good is the quality of your investment product data?

In one of my previous posts on Data Quality Denial I promised to follow-up with a post on how to evaluate the quality of your data…in fact when I am asked by someone from the industry “what do you?” – I often respond immediately with the question “How good is the quality of your investment product data?”.

Generally, you get a look of bamboozlement, because this is a question not easily answered. So the answer to the “what do you do?” is that I help people answer the question “how good is the quality of your investment product data?”.

So how do you set yourself up to answer this question?

First off the answer is completely dependent on your quality goals and objectives as stated by your data governance. If you do not have governance or any formal strategy around data quality management then the answer to the question is immediately intractable.

The question– how good is the quality… – immediately identifies that the answer is a qualification of quality – I am looking for a level of measurement…

Unless as an organization you have specified measurable targets for data quality then you cannot answer this question with any form of response other then “…ehhhh, good I guess…” or “…no one is complaining…”, or worst case “…we know it’s bad because the regulator just hammered us for X million..”, or “….we just lost a big mandate because of bad data in a report…”.

So how do you set measurable goals? In an earlier post on Data Damnation Part 1 and Part 2 – I mentioned the use SMART or SMARTER goals.

If you have strategic goals that fit the SMART bill then you can easily build the Key Performance Indicators or Balanced Scorecard to report on the exact quality levels your organization is achieving relative to the stated strategic goals laid out by your governance strategy.

If through good governance, you do have smart quality objectives then data stewards will be more capable of aligning business-as-usual operations with strategic goals.

The transparency delivered by having data quality measures reported centrally and visible to all – drives successful oversight and accountability. In addition SMARTER goals aid implementation of Six Sigma or ISO 9000 quality management systems through corrective feedback loops.

So what should be measured in terms of data quality?

Now this is the million dollar question because the answer really depends on who is asking it and what they want to achieve?

As you will appreciate the view of a marketing executive and a data source owner will differ wildly on the answer to this question. So each business unit needs to feed into the target and measurement process to ensure their specific needs and goals are being addressed by the governance team.

Each unit involved will generally be looking at 4 key measures – with specific twists for their own specific areas of interest:

– Accuracy: is the data correct?

– Timeliness: is the data late?, is the data stale?

– Consistency: is the data consistent across all inputs? is the data consistent across all outputs?

– Completeness: do all records have the expected attributes in complete form?