17 May by Stephen Boane
Longer article than usual about “managing the machine.” Hard to shorten it down. Stuff that interconnects, interconnects. SLB
Businesses do not work from a simple formula such as input plus process equals output. There is so much more complexity to any human endeavor and so much more variability. This is why decisions around performance measurement and tying that to compensation (including incentive compensation) are fraught with difficulty. Because businesses report in terms of money and volume it would appear that most things can be reduced to math. In some respects that is true, but it is not effective to be that limiting. Math gets you only part of the way down the road to create processes, measure performance, communicate results, motivate and incentivize results, and motivate the right behaviors that cannot be measured and cannot be incentivized. If it was easy it would be a lot less interesting.
When you read the advice around measurements and metrics it usually can be summed up with a statement like this, “Describe what is important to your business so you can measure it.” I disagree with this approach for four reasons:
- If you did a poll, most executives would measure broad indicators like revenue growth or customer satisfaction. There is not a thing in the world wrong with these when you go to write your shareholder letter and they may be just right at the most senior levels of the company. What they do not do is provide actionable information that your enterprise can use to react to its performance and the environment where is conducts business.
- Creating a measurement set based on what you think is important creates a bias towards measuring what you already know you can measure or based on intuition rather than analysis. There is a term called the McNamara fallacy which states that what cannot be measured is not important – see the previous sentence (!). Most of the time there are drivers that give you much deeper insight into your business and actionable direction for its improvement.
- If you want to scale and remain profitable then everything or nearly everything needs to be driven by consistent processes and good decision-making up and down the org. In even the most organized business most of the margin erosion and lost opportunity comes from places where people are still winging it rather than from process exceptions. This means rather than measure performance (what you think is important) measure process compliance and process effectiveness. If your metrics are positive and results are not where they should be, you are going to have to revisit how you are looking at your business. If your metrics are positive and the business is performing, then you have dialed in your performance measurement system just right (similarly if your metrics are negative and the business is not performing, you have valuable information).
- Last, by measuring what you think is important and using only that lens you are guaranteed to have an organization that does exactly what you ask for which can be a problem. If you measure customer count growth, they will find a way to increase accounts a la Wells Fargo. If you measure revenue growth you are likely to see a break down of pricing discipline or have a cost problem as people punch as much through the sales goal without considering other factors. Lets be charitable and say no one is being dishonest, maybe just disingenuous, and they are giving you exactly what you asked for. Kind of like seals, clapping for fish. Not at all like business people making solid, contextually appropriate commercial decisions.
So as the CEO, or President, or other C-level executive how do you think about performance measurement? How do you tie those metrics to comp or incentive comp? How do you avoid creating a maximization issue where making a goal does damage elsewhere? I suggest you look at it within a continuum that goes like this:
Core commercial model > Business processes > Process waypoints > Measurements & standards > Incentives > Desired outcomes
Core Commercial Model. The core commercial model is what the business does. Why exactly are you in business? For most companies, the core model is the customer(s) they serve, the product or service (or both) they produce, the mission and values they promulgate, the kinds of professionals they want to attract, the assets needed to get all that done. More than likely you have a good sense of what your core business is and how you define your economic model.
Business processes. The business has a series of business processes or activities that cause the core commercial model to occur. You can go really deep here and I encourage you to do so. At one company I operated, our field service process had a checklist that began with “Two sharpened Number 2 pencils in your company branded computer bag.” Start your process mapping with an event, like “Customer signs contract” and go through to an end like, “Collection of receivable.” Inside that wide space are all of the people and activities that go into your revenue model. This, then that. Building these processes is a company wide effort. If your looking at receivables, one or more of your Accounts Receivable Clerks should be on that process team. This means you learn exactly what is happening and if the mapping changes something, they are part of the change. Change adoption like magic!
Process waypoints. As you map out your processes there will be process waypoints that emerge. These are obvious inflection points where something critical has to happen or other stuff will not happen or could happen incorrectly. Most of the time are “handoffs” from one department to another, one location to another, or one system to another. These are places where an “internal customer/internal supplier” relationship is created. Back to our example, “Customer signs contract,” one waypoint in the process map might be this: “Clear copy of contract delivered to production supervisor for scheduling.” The salesperson is the supplier and the production supervisor is the customer. Production cannot occur without this and assuming time is of the essence, it needs to happen fast. Thus this action becomes a waypoint. Everything else in production happens because this occurred. This is particularly relevant if your process mapping team lets you know that the business has been experiencing a problem here. For instance: the salesperson emails the contract after they scan it but they are irregular in doing so; the documents come over in a format that is hard for the production supervisor to convert into production specs; there is no way of acknowledging that it has been received. The list goes on.
Measurements. Now we know there is an internal customer/supplier relationship between the salesperson and the production supervisor. The salesperson is producing a product, the contract, that the production supervisor consumes to schedule production. Since this is a critical handoff we can create a measurement and, as importantly, we can set a standard. Here are some suggestions: % of contracts delivered to production supervisor by next business day (standard: 90%); cycle time from contract signing to production scheduling (standard: 2 days); # of contracts received with product ready terms/# of contracts sold (standard: 100%). The list goes on. What each of these does is create a pragmatic conversation around performance; you likely only need one, maybe two at each important Process waypoint. As a senior leader how do you handle all of this? You want measurements and standards for every position, managers are talking to their reports about what is important and being measured in those positions, you are talking to these managers about performance using metrics appropriate to your relationship with them. These will be different by position because each position is responsible for something different. You do not need to know all of the measurements in the organization as you have your own that you are responsible for and you have specific measurements to discuss with the folks that report to you. When you cycle through a deep dive of a functional area at say a quarterly departmental review, your managers can go over what is happening at a more granular level so you can see the connections between performance deeper in the organization. More importantly you can confirm that your management team is seeing those connections also.
Incentives. Most of the time incentives will come right out of your Measurements and your Process waypoints. Incentives are based on a subset of what you are measuring. If you have identified five measurements in a process, you find the one that is best suited to an incentive and use that for a given position. Everyone ought to have an incentive because every position has a performance expectation that they are resourced to exceed. None of the incentives need to be large. Here are the key learnings I have had on incentive design, some obvious, some less so:
- In designing your incentive look at the benefit from overperformance against the cost of the incentive. This is true over time as your processes change.
- The biggest incentive of all is consistent discussion and recognition. Yes, more money is more money, but that only goes so far. Recognizing performance is much more important. In the case of our salespeople and production supervisor, I would shoot out a weekly email to the sales team with everyone listed and what their prior week and prior month measurements looked like. Seeing how they rank against peers is incentive enough on this one.
- Incentives tied to things that the person does not control are basically useless. If the company has a good quarter, great. If you pay your machine operator $100 that month because of it, he will be happy when it happens, dissatisfied when it does not. The incentive will not drive any particular behavior on his part but it will create a temporary reaction unrelated to his actual job. Pay him a $100 bonus every month that his production within spec exceeds 97% because you can draw a straight line from that standard to overall company performance.
- Be selective in how many incentives a position has. In most instances a weekly or monthly incentive plus a quarterly incentive on something else is plenty. More than that and the portfolio effect takes over and nobody can tell what is really important.
- The closer in time the incentive occurs to the activity, the more effective it will be. For folks doing production work a weekly incentive is fine although you may want to go monthly to reduce accounting complexity and make the individual dollar amounts more noticeable. Other incentives can be quarterly for managers for instance.
- Set an expectation that incentives and standards will change over time or time bound the incentive as a temporary initiative. This avoids creeping incentives that get out of synch with what a position should be paid. You can assume that your company is going to get better at what it does over time, that it will invest in productivity enhancing resources. When those things happen the incentive is not driving behavior, it is rewarding longevity. This is particularly true on incentives tied to financial measures like total revenues rather than to a standard like revenues over budget.
Desired outcomes. So we know what the company does. We know how it does it through process mapping. We have identified process waypoints and acknowledged points where we have internal customers and suppliers. We are measuring key activities and key handoffs and communicating those measurements to the right places. We are incentivizing a portion of those measurements where they drive performance critical to a process. All of this should be leading to increased performance in balanced way. If it is not, you have the tools at hand to go into the business and figure out what is not working. You also have a way of structuring your own time because just setting this up is not enough. You will need to go over this with your direct reports and pretty much everyone else you see. One of the very best operators I know owned a huge chain of convenience stores. Everything was measured. When he did store visits he studied the metrics the store managers used at the locations he was going to. I watched him thank individual store managers for individual things they did particularly well – metrics he did not follow daily but he knew they did. That is a tough business with a lot of moving parts since there were so many locations and so many people. He paid attention. He recognized that every job was important. He made exceptional performance look effortless. You can do the same thing.