Many times, our customers ask about how they forecast or how they model the impact of changes to their compensation plans or to their commission schemes for the new year.
It seems like a simple request, but the issue is a lot of assumptions are made about what modeling is, what sort of scenarios should be run or what underpins the modeling.
Learn How To Model Sales Compensation Plans
Today’s webinar is about trying to clear up some of those assumptions or misunderstandings about this space. In this video on sales compensation forecasting and modeling, Performio Founder, David Marshall is joined by TransUnion Senior Director SP Dave Egloff. Dave has led sales compensation at four different organizations in four different industries, so his perspective on sales compensation is certainly well-rounded.
1. Why Is Sales Compensation Forecasting and Modeling So Difficult?
It’s difficult because commission expense is a variable that is dependent on many factors, some of them being a coverage model, plan design, quota setting, and sales results.
2. What’s Wrong With the “Traditional” Model Approach?
It just has too narrow a set of assumptions. If the assumptions are wrong, and your calculations are right, you’re not going to have a good financial predictable value.
3. What Is a Monte Carlo Simulation?
It is a model that uses statistically relevant but variable numbers to predict outcomes over thousands of instances, simulated over many years.
4. What Is Sensitivity and Volatility?
Sensitivity is the variability that is caused as macro performance changes, and volatility is the spread of expense outcomes at a singular point of macro performance.
5. How Will the Sales Results Approach Be Different Now?
Armed now with Monte Carlo results, your executive comes to you and asks, “I love your plan design change proposal, but how much is it going to cost?”
You can say:
“If we hit our plan number, we’re going to pay relatively the same amount. But here’s a range of values” — and now you could even provide a lot more thought behind it. “If we miss our plan number by 2%, we’ll pay about three hundred thousand dollars less, plus or minus 3%. If we beat our number, you can expect to pay this amount more. But just so you know, I compared it to last year, and it’s relatively cost neutral to last year as well.”
That’s a far more compelling answer.
False Assumptions in Sales Compensation Modeling
So how many of us have been in a situation where one of our executives says, “Hey, I love your comp plan design change proposal. How much is it going to cost me, though?” As the good comp planners, we are, we often say, “It depends.”
Of course, it depends. And often, our forecasting models are based on the previous year.
However, will next year’s results be the same as last year’s?
That answer, obviously, is no. And right here we have the start of the faulty assumptions that go into modeling. I often describe an individual plan year as a fingerprint or a snowflake, in that there are so many idiosyncrasies that can alter commission expense. If you have a large sales force or you can sell multiple products or services, the complexity of the uniqueness of that plan year, or that fingerprint or snowflake, becomes even more difficult to repeat.
As we know, if we use faulty assumptions in modeling, your model is not going to be a good indicator of the future result. Your actuals will be different from your estimates, and you check your math — but the challenge is likely not in your math skills.
My guess is that you are a pretty good comp plan designer, and you’ve been doing this for a while. Your Excel skills are also likely pretty good. Save any minor formula errors — which you hopefully have validated that are not the case — you check your math. You assume that the math is correct, and you have the scrutiny of people looking over you.
There’s a better way of doing things, and it starts without assumptions. If you have a very thorough conversation at the onset, you can have thorough and comprehensive conversations as you go on through.
The Challenge of Modeling Sales Performance
So why is modeling so difficult? Commission expense is a moving target, and it’s a tremendously dependent variable. I’m sure some statistician is going to yell at me for saying “tremendously dependent.” It is a dependent variable, and it also goes into part of the explanation about why you must make certain assumptions, or why certain assumptions could be very misleading.
Let’s cover the dependency that commission expense has on some of the more independent variables. This is just an example — there are four broad strokes of independent variables, and some of the bullets are just examples of the assumptions and the contingencies that nest within.
A coverage model goes into your commission expense. What is your go-to market strategy? What is your territory planning, and are you optimizing your territories? How many credits per sales transaction are you going to offer?
If you have an organization that has a lot of overlays, a lot of management hierarchy levels, you’re going to have more transactions — and more credits per given sales transaction. Very often, this comes down to territory alignment and go-to market strategy, which is usually dictated by the sales management. Most of us are not sales management. This goes to show that part of this dependency is based upon an independent variable we don’t control.
Effective Sales Comp Plan Design & Quota Setting
The next thing to think through is plan design. This goes down to the plan dynamics: Is there a performance threshold in place? What is your upside potential? Do you have regressive pay rates? Is there a cap?
This is the sales comp designer’s role, and this is often us. So this is an area where we have direct influence.
Obviously, then you have is quota setting. Now, this is one of those that could be us — it’s probably not, but it depends on where our role sits. I’ve had quota setting responsibilities when I’ve sat in finance. I lost that when I went over to HR, and now I have somewhat of a dotted line accountability being in sales operations.
But often, when we look at quota setting, there are several important questions to be asked:
- First, are we good or bad at quota setting?
- Are we using revenue or a bookings metric?
- Do you use team-based targets?
- What is the impact of team-based targets versus individual targets?
- Is it new revenue versus a recurring revenue stream — or some sort of payment on an annuity stream?
These are very impactful to commission expense. Again, if we have influence that’s great, but oftentimes, this is not us.
The Importance Of Distributive Analysis
Finally, the last question we want to consider is what are the sales results. Sales results are the single accountability from sales themselves. How well did the sales team do against their targets? Did you have a bunch of extreme performers, or did everyone huddle around a hundred percent? What actual products did they sell?
Obviously, this is not us, and I can’t underestimate or understate the impact of distributive analysis.
Let’s use an extreme example. Imagine you have an organization where you have a decent number of salespeople. They all have the same quotas, and they all sell right at 100 percent. That’s one scenario. In a parallel universe, you have the same salesforce, but instead, you have half of them sell at 90 percent and the other half sell at 110 percent.
Your commission expense will be very different, even though your average attainment is 100 percent. That’s often because your over-retainers will earn more than the under-retainers will save you. So that’s why it’s important to examine the distributive analysis.
The point here is that when you build a model, you must realize you’re going to make assumptions with all these things in mind, and most of these things are out of your span of control.
Various Options for Sales Comp Models
When I was first coming up the ranks of sales compensation, I did not have very good mentoring and coaching.
I was told that the best way to go model was just to take the plan design and put it over last year’s results — and if you want to be really thorough and impress people, why don’t you do it over the last two or three years?
As you know, this isn’t a very good calculator of the future. What you’re essentially doing is you’re taking the new plan, comparing it to those years, and seeing how the plan design would change.
This method does not provide an effective measure or assessment on sensitivity or volatility, and it doesn’t give you a very good comprehensive story on the plan design dynamics.
And honestly, if we can all agree that it’s not a fair prediction of the future result, what are we doing in the first place? We’re taking a few bits of anecdote (and a given year might as well be an anecdote!) and we’re trying to draw broad conclusions about what the unknown is going to bring us. It just does not work.
Welcome to my Captain Obvious moment. Commissions are a very large expense, and it’s a very variable expense.
Because of this, you need to look at comments like the traditional approach not being a fair predictor of the future, and really challenge if you’re going about it in the right structure. The first question you often get is, “Well what would it have done last year?” This is a good opportunity for you to show off how well you know your sales comp design, and how good you are by being able to propose an alternative method.
The Alternative Sales Comp Method: The Monte Carlo Simulator
Let me introduce the Monte Carlo Simulator. A Monte Carlo Simulator is a model that uses random but statistically relevant variables. It allows you to repeat a scenario with random numbers, so you can see how a plan design performs in a multi-year simulation — and it’s not last year’s simulation.
It is a probable year’s simulation.
You can repeat this exercise, and I tend to repeat it for 20 years. I feel like that’s a number that communicates well, and I’ve done it for 25 years with no material change. Now, I can show you how this plan design will work in 20 random years, as opposed to just telling you what it would have done last year.
Because the approach is far more robust and far more comprehensive, the predictive ability of this model is far higher than any of the traditional approaches. The inputs you’ll need are probably things you have readily available.
For example, you need your plan designed by role. You need your sales attainments or your quota attainments by role — meaning your historical averages and standard deviation of quota attainment. If you do not have a plan design that uses quota attainment, you could use the revenue or bookings absolute dollars as well, and then target incentive by role.
There are two ways to go about target incentive, or your target commission, or your variable pay — depending on where you are, the metric is called different things:
If you wanted to predict what you’ll pay next year, I would suggest that you age it forward and apply merit and salary increase, if your sales force is eligible for that.
If you just want to compare an apples-to-apples model, where you’re trying to express the cost impact, then you can use the current target incentive by role.
If you take this approach, the insights you will grab from it will greatly change your approach. It’ll change the conversation with your executive leadership, and you’ll not only be able to model in a better fashion, but you’ll be able to draw insights about your plan design in a much better fashion.
Here is an example of what I provide in terms of a visual diagram for the Monte Carlo Simulator. I do it in Excel:
- First, I gather my statistics on sales payment targets and actual pay. These are easy numbers to grab if you don’t have them already, and you can use very simple formulas in Excel. You can figure your simulator, your Excel document, to the comp plan mechanics. This should be the equivalent of building a row that calculates someone’s target, or someone’s commission based on the target commission in their quota. This is something we often tend to be good at as sales comp practitioners. You can repeat these rows as many times as you want, and each row or each trial would represent a given person.
- You can then put this in a ‘what if’ analysis to repeat those lines. Each one of those has their own set of randomly generated numbers. Each year looks different from the previous year, and that’s really where the power is. It’s the repeatability of randomized years.
The reason why I made sure that I could do this in Excel is so that if I wanted to send this to someone, I didn’t have to send them my credentials to a fancy tool. Some of the latest and greatest tools have a few limitations on their modeling capabilities, so I wanted to build a tool that was system agnostic. And I feel like this is the right way to go about it.
This chart is very, very powerful. I not only use it for myself and with finance, but I also use it with our executive leaders because it really does tell a comprehensive story:
- The first thing I do is aggregate attainments on the bottom. I put three here: 95%, 100%, and 105% — and I could also do five of these if I wanted to do more increments, or I could do narrower increments. Most of the reason why I’ve done this is that I’ve gotten the question that said, “Okay, if we do one percent better than planned, what does that look like in terms of commission expense?” This helps answer that question. So those are highly customizable.
- The next thing I do is edit a referential point. That referential line can either be what we’ve historically paid in terms of commission expense, or it could be our target commission expense. Again, that is adjustable as well.
What Is Sensitivity?
Let’s talk about sensitivity — I’ve mentioned it a couple of times now. Sensitivity is the change in commission expense as your macro attainment changes. The question of sensitivity comes down to this: If I have one more percent of attainment, how does that impact my commission expense? Or subsequently, if we miss plan by one percent, we miss our revenue plan by one percent. How much will we save in our expense plan?
These are powerful questions you should be able to answer with confidence, and a tool like this will allow you to do that. In this case, I’m illustrating the commission expense change as we go from 95% to 100%.
What Is Volatility?
I get this question quite often, “Wow, why is that body so fat?” Or “You’re really telling me that we can have a two-million-dollar variance?” And I say yes. The reason goes back to your plan design. This is your probable pay range at a given attainment.
So if you have a lot of volatility or a lot of spread between your lowest probable and your highest probable payout, it is because there’s something in your plan design that is causing that bit of risk. Maybe your plan has high upside. Maybe your salesforce has very different target commissions. Maybe you do a very poor job of quota-setting, and to get to 105%, you might have some really high attainers balanced out with some very low attainers.
If you see this and that is the distraction for your executive team, you can share a couple of ways to fix it:
- One, you can do it through quota-setting. You can improve your quota-setting ability, which is far easier said than done. But you can attempt to fix it through doing a better job of setting quotas.
- You can also do it with comp plan design. You can do things like level out the upside earnings potential so it’s not as steep. Or you can put a regressive pay rate in there, or you can put a commissions cap.
I’m not necessarily endorsing those plan designs, but if my executive team has a lot of heartache with the volatility that our comp plan design has, I can offer them suggestions.
And this volatility also can contribute to accrual issues. So if you’re an organization that is struggling with accrual, or you’re depending on your financial accounting processes, it might be something you can visualize in this format. Say, “I have a better way we can do it. Here’s part of the solution, and here are some opportunities for improvement.”
That chart is remarkably powerful for anyone who has both a numbers orientation or a visual graphic orientation.
Benefits of Monte Carlo Simulation
So here are some of the benefits of the Monte Carlo Simulation. It is a far more robust examination of plan design.
We’re looking at 60 years of simulated results, as opposed to the last one, two or three years. The result set is just far richer than the traditional approach. We can also now both describe and graphically illustrate sensitivity and volatility. Remember, sensitivity is how your expense changes as your macro attainment changes and volatility are the range of probable payouts at a specific average attainment. These are not things you could have done with the traditional approach.
And again, the whole reason why we do this plan design, or one of the reasons, is improving your financial predictability. The other reason you do it is to have more insightful conversations around plan design. You are improving your financial predictability, and you are now having the ability to have a more insightful conversation. To me, those are game-changers.
We talked about sales commissions being a very high variable expense. My question is always, “If it is one of your largest expenses, and one of your most variable expenses, can you afford not to be doing something more robust?”
For my organizations, I’ve concluded that we must do a more robust method. It’s what I owe the organizations, being the financial steward of the commission expense.
After this conversation with your executive, the only thing for him to say is how great you are, how smart you are and how much you deserve a promotion.
Well, you may not get the promotion — but I promise you that you’re going to get people to look at you in a way that they would not have looked at your predecessors. Because that’s how much information you’re going to be able to provide to them.