Are you Collecting Bad Print Shop Metrics?

This is the first article in two-part series about issues with metrics. While metrics and analytics are important to manage your operations, they are not without issues. Two issues are bad metrics and the manual collection of measurements. For the last few months, we have discussed the importance of metrics and analytics to run your facility. One example was the blog post that almost half of in-plants collect operational data manually. But no conversation about this subject would be complete without discussing the issues.

Not everything measured is a worthwhile metric; some are bad metrics and lead to bad decisions and bad strategies. One example is when Wells Fargo offered incentives for account reps to open "new accounts", resulting in opening 3.5 million credit card accounts without customer consent. Trying to motivate staff using the new accounts metric was a disaster.

Popular Metrics, Vanity Metrics and Suspect Metrics

Two popular metrics used in print production are the number of number of impressions and the number of pages inserted in envelopes. However, without a measure of rework or accuracy these are bad metrics. Without a quality metric this may be measuring rework, such as the pages reprinted or reinserted.

Vanity metrics are measurements reported to make you look good, but they can be bad metrics. For example, when tracking online activity some people track registered users, page views and downloads. Believe it or not a vanity metric in printing could be on-time delivery, because the method of calculating on-time delivery can change though the course of the job. For example, when something goes wrong during a job the client may be contacted and asked to reconsider the delivery date. Using renegotiated delivery dates can result in 95-98% on-time deliveries but customers may be dissatisfied, leading them to request shorter turnaround times or "pad" their due dates because they expect you will call them and ask for more time.

Suspect metrics are bad, too. Years ago if you shopped in bookstores and started to read books you would be asked to leave because the thinking at the time was that people who read books in the store would not buy the book. Shopping for books is much different today. In addition to less bookstores, you can download and read samples online and now customers are encouraged to read books in the store. Most stores have comfortable seats, coffee, and food available. If you tell a bookstore or online manager that the people reading those books were not likely to buy the books they would say that is suspect information.

Almost every in-plant print center collects some data they consider suspect. One measurement often considered suspicious is job cost estimates. In fact, it is not just the estimate, but the building blocks of the estimate are suspect too, meaning the budgeted hourly rate or the time estimate. The evidence is obvious because either the estimator or manager reviews each line item and inevitably find something suspicious and say, "That is not correct" and then change it.

The more important question is why use a metric or an analysis that is suspect? It takes much more time to constantly check and change it than to simply fix it. This is a good example of working as a firefighter instead or fixing the process and employing fire prevention.

Conclusion

Just because you can measure something does not mean you should measure it, because some metrics are bad and not worth measuring. The decision about what to measure and analyze comes down to a simple cost – benefit analysis. How much does it cost in terms of time and effort and how does the collection and analysis of those metrics benefit the business? If it is a vanity metric, suspect metric or any other bad metric it's value may be small.

One strategy to consider is pretend you were just hired as a consultant and deciding what metrics worked and didn't work. You would likely ask "How is this measurement helping you to improve your business?". If a metric directly contributes to improvement, then it is a good metric. If it does not directly help you improve then it may be a bad metric.