Revenue Uncertainty – Part III: How to Model Revenue Risk

Tension is high, and anticipation is thick as the annual sales kickoff for DisruptaCorp begins. Employees at the young tech startup settle in their seats. Mobile devices are hurriedly silenced and stowed. Chatter dissolves into quiet.

The CEO, Priya Neghandi, stands in front of the room alongside her VP of Business Development, Kelvin Wickersham. Without saying a word, Priya takes a black marker and scrawls a single number at an upward tilt on a whiteboard, and swiftly underlines it with a confident stroke.

$15,000,000

She stops for a moment, turns, and gazes across the noiseless room. “That is our sales goal for next year,” she states calmly. Her resolute demeanor is infectious. Kelvin looks directly at his team sitting in the front row, and exclaims, “We have our goal. Let’s go take that hill!” Cheers erupt. Hugs and fist bumps all around. Everyone feels confidence and love. Life is good. But nobody breaks out the bubbly. Not yet, anyway.

In sales, this is a common vignette, emblematic of a deterministic approach to goal-setting. A senior executive or committee establishes a single numerical target which is indelicately lowered onto the waiting shoulders of the business development team. Assumptions and logic about how the target was derived are not discussed. Shaky words such as might not, could, should, probably, maybe, and likely are notably absent. All conversation centers on how to achieve the goal. “We should do webinars!” “We need the right content, and we’ll get the sales force trained on how to use it!” Few, if any, talk about what could derail their efforts. “I want to know how you are going to make your number,” Kelvin growls in his team meeting immediately afterward, “not how you aren’t . . .”

But, what if DisruptaCorp moved away from determinism? What if Priya hedged a little, and candidly admitted that she’s, well . . . not completely sure about achieving the $15 million target? What if she confided that there are things she doesn’t know? For example, whether the expected number of customers will upgrade to the new software release, what happens if the operating platform that Development plans to use is late to market, what will occur if the company’s main competitor introduces a better product three months ahead of forecast, and what will happen to demand if the economic recession deepens?

How would that candor change conversations? Which new insights could be revealed? Which actions would be taken?

Replay the kickoff meeting scenario, except now, imagine that Priya approaches her company’s revenue challenge differently, and writes this on the white board:

2015 revenue target –

Worst case: $7,000,000

Most likely: $10,000,000

Best case: $15,000,000

By uncloaking her reservations, Priya has initiated a key step for managing revenue risk. She has suggested that there are uncertain conditions – namely, forces and events – that can cause different outcomes. Her team begins to think about future situations, and the pressures they will exert on the company’s revenue strategy and tactics. They begin to think about their likelihoods and the ways they could converge.

Instead of instinctively running pell-mell toward The Hill, Priya and her team have undergone a paradigm shift in their worldview. No less tenacious and focused, they now have situational awareness. DisruptaCorp can begin to evaluate the future in terms of probability and risk, and the company can determine what matters. Most important, Priya’s team can anticipate trouble, and can take action before the revenue graph takes a southward turn. Similarly, they can also recognize positive forces and developments, and be in a better position to capitalize on the opportunities.

Priya’s openness helps the conversation grow and blossom in new, productive directions. Is the most likely revenue scenario of $10 million too conservative? Is it fair to assume that no particular revenue outcome is likelier than any other – making most likely simply the average between worst and best ($11 million)? And, what about the probability of hitting equally important revenue targets, such as break-even, which DisruptaCorp’s CFO has pegged at about $8.8 million? Without further analysis, it’s impossible to provide intelligent answers. But at least now the questions have been raised!

By considering worst case-most likely-best case, DisruptaCorp’s team has also discovered new questions to ask, including:

– How much to invest in lead generation investment to cover expected customer churn

– Whether hiring additional salespeople will reduce the risk of achieving their revenue target

– What is the best price point when entering a new market, like healthcare

– Whether offering volume discounts will improve net revenue

– Whether investing in skills training and staff development for the sales force will reduce the probability of missing break-even revenue

For revenue planning, positive correlations between variables are not difficult to identify, explain, or understand. More lead generation effort generally leads to more revenue (though not always efficiently). Greater social media presence increases the opportunities to converse with customers online.

But other relationships are convoluted. Price increases don’t always result in greater total revenue. Reductions in defect rates don’t always improve customer loyalty metrics. Plus, situations can combine in millions of different ways, which makes revenue planning a Sisyphean task. How do executives align investment and effort to needed outcomes? With so many pieces and parts, it’s nearly impossible to manage day-to-day operations without the aid of statistical risk analysis.

What underlies revenue volatility are uncertainties and risks that have come home to roost. The hot product that a competitor just launched with a big media splash. The top sales producer who quits without warning, taking his biggest accounts with him. The customer complaint video that embarrassingly went viral. The corporate Tweet that overstepped the boundary of good taste. These situations underscore why determinism – anchoring on revenue goals without accounting for risk – creates failure.

Fortunately, for business planning, many uncertainties can be accounted for because they can be modeled and analyzed probabilistically. In my previous article, Part II: Putting Uncertainty to Work at Your Company, I outlined five steps for exposing risks that jeopardize revenue. This article explains how to use statistical models and Monte Carlo analysis to develop a more realistic vision for revenue achievement under a set of assumptions or conditions.

Here are the next five steps:

1. Select a distribution model for the variables that are consequential for revenue. Examples include unit price, cost, demand, and currency valuation. These models will be used for statistical analysis to determine outcomes of interest for planning targets – for example, the expected cost for achieving different pipeline multipliers, how many new customers to acquire for achieving 25% revenue growth, and how many new salespeople to hire. The top question to ask: based on the volatility inherent in the ranges developed in Step #5 in my previous article, which risks are low-risk risks, and which are high-risk risks? The answer to this question enables companies to prioritize what’s considered for statistical risk analysis.

2. Take a “reality check” on the expected distribution by asking “does this appear right?” If not, adjust the values.

3. Run scenarios that randomly combine key variables using Monte Carlo simulation. In the example that follows, we can examine how different probability distributions for cost, price, and demand might combine.

4. Develop quantities of interest that can be modeled by establishing a relationship between the variables. For example, gross revenue (derived by demand times price), net profit (price minus cost), customer service staff level (the sum of all inbound contact incidents divided by number of inquiries each agent can handle), etc.

5. Discover new questions to ask. Once the key variables have been identified and the best probability distributions have been selected, decision makers can ask many other planning questions. For example, based on future assumptions, what is the probability of achieving $200 million in revenue in year three? Or, if DisruptaCorp loses money in a given year, what will be the average expected loss?

Remember Priya’s original goal of $15 million in annual revenue next fiscal year? The statistical analysis provides a nuanced perspective. She will probably fall short by about $4.5 million, based on her team’s worst case-most likely-best case estimates for customer demand, cost, and selling price (see Table 1).

I discovered that after making 10,000 stochastic (random) iterations of the probability distributions for the Table 1 variables (see Table 2). But, there’s also some good news: 1) DisruptaCorp will likely be ahead of break-even by about $1.6 million, and 2) their projected profit margin is 34% – well above the CFO’s target of 25%. But best of all, DisruptaCorp can determine these likely results beforehand, and take action to manage the risks.

Next year at this time, Priya and Kelvin want to open a case of champagne for the DisruptaCorp team to celebrate a sales year in which they achieved (or over-achieved) their $15 million goal. Considering the estimates they have provided in Table 1, what is the probability of that happening? The Monte Carlo analysis tells us that 29% of the time, the variables will align to produce that outcome.

Priya must keep her investors happy, and she clearly wants to improve the odds. Through risk analysis, she can explore the most effective way to ensure that. Priya believes that if she she can reduce demand volatility (now between 1.5 million units and 3.7 million) by beefing up outbound marketing, DisruptaCorp will mitigate some risk in achieving its revenue goal. But will the risk reduction justify the cost? And, if selling price does not drop below $4.00 per unit, will that improve overall revenue, given the risk that it might also reduce demand if the economy does not improve? With risk models, Priya’s team can test different scenarios and develop the best strategies and tactics to meet DisruptaCorp’s objectives.

Table 1 – Ranges of key variables influencing revenue

	Worst case	Most Likely	Best Case
Expected demand (units)	1,500,000	2,500,000	3,700,000
Price per unit	$3.80	$4.00	$4.75
Cost per unit	$3.15	$2.80	$2.15

Table 2 – Variables after 10,000 stochastic iterations:

	Key variables after simulation:
Expected demand (units)	2,612,546
Price per unit	$4.01
Cost per unit	$2.65
Total Revenue	$10,477,112
Profit Margin	34%

Compared to determinism, probability analysis adds many new complexities. After all, what could be easier at the outset than pointing toward a revenue hill and encouraging your team to go take it? But risk modeling will help you figure out whether there are any obstacles in between, and it will enable you to understand and estimate their magnitude. That insight will help you get past them, better ensuring your success.

Which champagne should DisruptaCorp buy before next year’s sales kickoff? No doubt, that’s a problem Priya and Kelvin will be delighted to have.