In this article we will work through a real example of a smooth operational KPI that changed a company’s relationship with its customers. After reading it, you will be able to apply a similar thought process to your own business.

Establish target service levels.

The business: specialized auto component repair services. The company had a KPI model in place.

The Customers: Auto repair companies that deal with consumers and their insurers.

The service: Collection of repair work, repair in a dedicated specialized facility and return to the customer.

The symptom: Frequent inquiries and constant complaints about the delivery date and time.

The real problem: jobs are queued in the order they are received, regardless of the complexity of the repair. Small, simple jobs would wait because they were queued behind large, complex jobs.

The first solution: Jobs scheduled to meet a specific day and time of return; complex jobs would be scheduled for a later time. The customer was informed of the execution of the return delivery with the quotation before starting the work.

It worked? Partially. There was no target service level or performance measure. It was not tied to the business marketing strategy.

Set your target service level: Ask customers. A quick and dirty phone survey suggested that 48 hours was an acceptable delivery time, because the bottom line was your ability to tell the customer when the vehicle would be ready.

Performance Objective: A 2-month history evaluation showed that about 15% of the jobs were complex and could not reasonably be completed in 48 hours. We assumed that 85% could be the correct point.

Target service level: 85% of jobs will be returned within 48 hours. This was a promise that no competitor could match.

Measure of performance: We predicted that if the target LOS was achieved, complaints and inquiries would drop to a low level. The complaint log showed that complaints were reduced to zero within 4 weeks after the new system was launched and advertised to customers.

The Results: Market share increased dramatically as the auto repair industry embraced the new standard and word spread that the service was reliable.

Internally, productivity increased due to better scheduling. Disruptive inquiries and demands for special treatment were virtually eliminated.

Operating profit increased.

Soft KPI test

On three occasions over the next 2 years, the complaints suddenly increased in volume. A quick review of the service level showed that it had dropped below 85%. Immediate corrective action restored the incidence of complaints to its normal low level.

My conclusion: we were fortunate that our first estimate of the desired level of service was correct. If customers needed 90%, then we would not have seen the desired reduction in complaints.

If customers had been happy with 75%, we could have experimented with that and monitored complaints. If they went up, we could go back to 85% LOS, but we would have unnecessarily annoyed some customers and put our goodwill at risk. Also, some competitors could have reached 75%, so there would be no competitive advantage.

85% LOS worked for everyone, so don’t play with a winning formula.

This is a classic “Soft KPI”.

Measuring it does not help you understand what is happening. Understanding what is happening does not help you find a solution without measurement. Both are necessary to make sound decisions.

Establishment of soft KPIs

Look for performance indicators that logically influence business performance, but where the effects are distant in time and place from the cause, and where it cannot be linked to your KPI model using an algorithm.

An example is job rotation. Everyone knows that this is very costly and that high turnover has very detrimental effects on business performance.

Is job rotation a KPI?

It cannot be a strict KPI because it can only estimate the effects on profitability with varying degrees of uncertainty. You cannot incorporate it into your accounting system even though you may find some of the costs eventually buried somewhere in your accounts. The rest of the costs come from the loss of something that drives a strong KPI, usually in the way that the loss of a customer leads to a drop in sales. The reason is that you cannot predict the cost of losing a single good employee, or the added benefit of losing a single bad employee.

It all depends, but on what?

You can correlate job turnover with profitability, but only over an extended period of time. You can’t define a formula that ties job turnover to a strong KPI with confidence.

Job turnover can be a soft KPI, but only in some companies. In highly seasonal businesses that rely on contingent labor, contingent recruitment and training costs are high, can be factored into the budget and managed as a stringent performance indicator.

Compare that to the loss of a key seller; The value of your knowledge of your business and your customers is always difficult to estimate and expensive to replace.

Therefore, a soft KPI must satisfy some, perhaps all, of the following criteria to be useful.

* Clear correlation of the soft with a related hard KPI.

* A metric that allows you to track performance against the soft KPI.

* Clear links to a controller KPI or “consequence” KPI that can be tracked. Driver KPIs are leading indicators.

* The ability to estimate a threshold that can operate.

* Is it mission critical? Are you warning of a potentially fatal error?

Can you fit them into a KPI model?

Probably not. That does not mean they are not “mission critical.” They must be tracked.

Product safety issues that drive product recalls are mission-critical, as Toyota recently discovered because of their high cost. I’m sure Toyota knows what their hard recall KPIs told them after the event, but the soft KPIs that might have highlighted the cause of the problem were clearly absent.

Threshold effects confuse the soft KPI target settings.

Many cause and effect relationships in business are not a linear correlation. In some cases, there is a threshold effect at work.

The problem of setting the right level of advertising spending is a good example. We know that in some markets the ad spend has to buy enough space to be visible to customers. Spending below that threshold, wherever you are, does not increase sales, because below the threshold you are invisible. Above the threshold, the results flow,

Likewise, looking at our service example, setting the target LOS too low would accomplish nothing in terms of customer satisfaction. Reaching the correct number worked like magic.

Places to look for soft KPIs

Quality and service are essential functions, in which mission-critical soft KPIs can be found. As I have shown, threshold effects are common. Both physical and soft KPIs are useful in evaluating product and service quality and it is productive to link the two types of KPIs.

Project management is equally dependent on soft KPIs and strict measures. The predictive element of project forecasts is inevitably soft because it uses probability measures; critical, but only tough after the event.

In the field of sales and marketing, measures such as customer satisfaction are confused with the challenge of finding reliable measures, and the correlations with strict measures are not as high as we would like. Market share figures are often dependent on industry statistics or surveys and can be quite unreliable.

Employee morale and engagement are clearly important and difficult to measure reliably because the inherent errors associated with opinion polls confuse the results.

You must be alert to the need for leading indicators. These are the most useful, because they allow you to take corrective action before things get out of lane.

The sales forecasts, indeed all forecasts, are soft KPIs, even though they are provided as hard numbers. They are KPIs, but every forecast should have confidence limits around it so that you can assess their reliability. Forecasts are, of course, mission critical in most companies.

Soft KPIs and their management process

Soft KPIs offer great value to your management process, so you should never ignore their existence. It’s just that the differences between soft and hard KPIs mean you need a different approach to how you develop them and how you use them. By following the advice and thought process illustrated in the story, you’ll be well on your way to identifying your soft KPIs and putting them to work.

Hard KPIs are easy to work with because they have one characteristic that doesn’t change; they are mathematically linked to measurable changes in business performance.

Soft KPIs are a bit more difficult, because you can identify the reasons why they are important to your processes, but there is no algorithm to define how they interact with business performance.

When you find the soft KPIs that are meaningful to your customers, you can refocus your staff on the things that really matter and transform your business.

Soft KPIs are company or sometimes industry specific, so you may find a source like kpilibrary.com helpful for generating ideas.