Payback Time: Evaluating Document Capture ROI

Posted by Mark Brousseau

Calculating return-on-investment (ROI) is all about the numbers, and the simpler the better. But be careful to focus on the right numbers or you risk missing process improvements that really count. KeyMark Inc.'s Brian Becker explains:

For any document capture system where you put automation in play – whether basic indexing, or complex document classification – simple recognition rates would seem to be the obvious measure. The reality, though, is that simple optical character recognition (OCR) and classification rates tell only part of the story, and often only a secondary part at that.

Ready. Aim…
Take a minute to list the goals for your document capture system – the factors specific to your business needs that will define success or failure of the project at the most basic level.

Now look over that list. Whether you listed two goals or a dozen, OCR recognition rate is likely not among them. That is because the real factors that drive your bottom-line ROI will be things such as:

... Efficiency improvements / Reduced manual labor
... Higher throughput (quicker availability of documents)
... Quality improvements (fewer errors = fewer corrections and improved customer experience)

While recognition rate does play a part, it tends to have little meaning on its own. It can be useful (specifically during system tuning), but it will always be secondary when measuring system results

It’s all about... You
The key to optimal analysis of your system is to identify the most important goals for your business. What business problem are you attempting to solve? What are the measurable parameters for your goals that indicate success?

For each area to be measured, what is your target? This might be a percentage reduction in number of errors, or maybe a specific number of documents-per-hour per operator. Use whatever makes sense in your environment for determining the benefit of the new system.

Is your goal to allow each of your operators to continue to classify and index your current document volume, but in half the time? Measure system throughput (per FTE-hour) and see how it stacks up to your old numbers.

Is your goal to reduce indexing errors by X percent? Analyze a sampling of work performed the old way, and compare the error rate to that from a similar sampling of work performed using the new system.

When complete, your list of goals and targets should be a solid basis (along with associated cost factors) for calculating the ROI, both planned and actual.

By the way, once you get your new system in place there is a good chance you will identify additional benefits, and will want to measure and compare these to the old process. While it may be beneficial to do so, be careful to keep your focus on areas of measurement that tie directly to your justification for the project. Since your goal is to measure actual vs. planned ROI, if you step outside of the original boundaries be sure to treat that separately.

What matters most
In the end, the success of the system will hinge on a solid design and implementation, with recognition rates playing an important, but limited part. A well-designed system can provide a great ROI even with mediocre recognition rates, whereas even the best recognition rates cannot overcome poor design or execution. It’s the total package that counts. (And, of course, you do want the best recognition rates on top of the best system!)

When you evaluate the success of your system, numbers don’t lie. Just be sure you’re looking at the full picture so you can be confident that you are measuring the parameters that tell the real story. Get it right and you will know exactly when you have achieved success.

What do you think? Post your comments below.