A Probabilistic Budgeting Framework | Jeff Epstein (Bessemer, Oracle)

Probabilistic “thinking” isn’t just for LLMs - it’s useful for budgeting too

Happy Thursday! This week’s post marks the beginning of a new type of Steelhead content - the Builder Series. In the Builder Series, I will interview founders and operators about topics they are experts in. The goal is to go deep rather than wide during these conversations, and then distill out the key, actionable insights for you. Given the depth and relative specificity involved, not every post will be relevant to every reader, but over time this content will become easily searchable so that topics that are relevant to you can be easily found. Let’s jump in.

This week I am thrilled to tell you about my discussion with Silicon Valley legend Jeff Epstein. Jeff has had a storied Silicon Valley career - first as an investment banker, then as a CFO for 25 years (including at a company you may have heard of, Oracle), and now as an Operating Partner at Bessemer Venture Partners. In addition to his full-time roles, Jeff has sat on and currently sits on many boards, including Booking Holdings, Twilio, Okta, Kaiser Permanente, and Shutterstock, to name a few.

As you might imagine, narrowing down my discussion with Jeff to just one topic was hard. Given Jeff’s experiences, he could write a book on any number of subjects. We ultimately decided on a topic that Jeff had previously written about - the process of building a budget.

I like this topic because it’s something that every entrepreneur beyond a certain amount of revenue (Jeff says about $5M) will have to face. Even if you are not yet at that point in your business, it’s best to understand the mental framework now for when it becomes necessary later.

Jeff wrote a great piece about budgeting - which he titled “The Goldilocks Budget” - on Bessemer’s blog. You can find that article here. We used this piece as the jumping-off point for our discussion. I’d encourage you to read it in addition to this post.

Jeff’s “Goldilocks” Budgeting Framework

Jeff became an expert in budgeting over his 25 years as a CFO and 30 years as a board member. In fact, it’s his time as a board member, one meeting in particular, that got him thinking about a better way to build and rally a company around a budget.

During that meeting, the chair of the board, the CEO, and the CFO all looked at the budget from different perspectives. It was clear that there wasn’t alignment on whether or not the goal was to hit the budget, or if under-delivering was a fine outcome because the budget was intentionally aspirational to push people to achieve greater outcomes.

In his article, Jeff mentions that there are three mindsets when setting a budget. These form the basis of how he came up with the name “The Goldilocks budget” - not too much, not too little, but “just right”.

• The Elon Musk / Papa Bear mindset - Aggressive. “In order to have a good outcome, we must strive for a great one.”

• The Charlie Munger / Mama Bear mindset - Conservative. “The first rule of a happy life is low expectations ... if you have unrealistic expectations, you’re going to be miserable all your life.”

• The Andy Grove / Baby Bear mindset - Balanced. “Objectives should be set at a point high enough so that even if the individual (or organization) pushes himself hard, he will still only have a fifty-fifty chance of making them.”

  

  DALL-E’s depiction of Goldilocks and the 3 Bears during budgeting season

Again, from Jeff’s article:

Every company is different, and the same company is different from one year to the next. There’s no universal formula for how aggressive or conservative to be in budgeting. The key is discussing it at the beginning of the process. The CEO can first lay out the vision for the year, discussing and selecting an aggressive, balanced, or conservative approach, in partnership with the CFO. Then they can discuss this approach with the board of directors. Not every director may agree; they can disagree and commit instead.

Getting down to brass tacks, there are 4 key components to a budget:

1. Revenue

2. Profit

3. Wall Street guidance

4. Sales quota

The key aspect of Jeff’s Goldilocks budgeting framework is thinking in terms of probabilities. During our discussion, I asked Jeff what his one piece of advice would be for someone new to this topic. His response: “Think about your budget in terms of probabilities”.

Next, it’s critical to make sure that the CEO, CFO, and Board (at minimum) are aligned on the methodology ahead of time to avoid unnecessary future board room frustrations.

In Jeff’s article, he recommends setting probabilities for each of the 4 budget components above as follows:

1. Revenue - 50%

   1. Jeff says that for revenue, he likes the Andy Grove approach. There is something about a 50/50 win rate that motivates teams to perform at their best.

2. Profit - 70%

   1. For profit, Jeff says that it’s important to be more conservative and protect the company from large misses on such an important metric (especially in this new non-ZIRP macro environment). He adds that it’s useful to add an expense line here for unexpected “contingencies” so that a miss in revenue may protect profit.

3. Wall Street guidance - 90%

   1. For those companies for whom this is applicable, it’s important to build credibility with Wall Street - therefore a 90% success rate is advisable.

4. Sales quota - 30%

   1. This one is interesting. Jeff mentions that sales quotas across companies are often optimistic. Given this, if you want to achieve a 50% revenue achievement rate, you need to build in a buffer. Jeff has found 30% to be a good target achievement rate.

Of course, there is no “right” answer when it comes to creating a budget. Every business is different, operates at a certain time in its lifecycle as a living entity, and operates in a changing macro environment.

For example, Jeff says that:

If you’re operating in a high risk environment, aim lower and target a higher probability of success. For example, you can follow this approach when:

• It’s hard to raise money.

• You have high customer or supplier concentration.

• You’ve missed your budget several years in a row.

• The economy is getting worse.

If you’re in a low risk environment, you can aim higher, be more ambitious, and accept a lower probability of achievement.

Once the budget is set, it should not change. When it comes to quarterly (or more frequent) forecasts, however, the same logic and probabilistic thinking apply. Importantly, when forecasts are made which necessarily include some actual data for the year (e.g. Q2 includes Q1 actuals), that new forecast should still maintain an X% probability of attainment.

Furthermore, in Jeff’s article, he mentions that there is one exception to the budget never changing. If after two quarters, it’s clear that the original budget is far off actual performance, throw it out and create a new budget for the second half of the year, using the same probabilistic thinking.

Overall, the goal of a budget is similar to the goal of OKRs. You want to set ambitious but achievable goals to push the organization to perform its best. The core insight to the Goldilocks budgeting framework is that thinking in terms of probabilities can help you get there.

The above was largely a summarized / paraphrased version of the article Jeff posted on Bessemer’s site, with some details removed. Again, I’d encourage you to read the full article for full context.

I wanted to use my time with Jeff to go deeper into a few topics that weren’t necessarily covered in his article; continue reading to learn more.

A Deeper Dive Into Budgeting With Jeff

During our conversation, Jeff and I explored a few other budgeting topics. In a Q&A-type format, you’ll find the questions I asked Jeff and his responses below.

What’s the best way for leaders to go about implementing a new budgeting framework at their company?

• Simple, it starts with the CEO and the CFO aligning on the new approach. Once alignment has been achieved between the CEO and CFO, it should be conveyed to the board so that all key stakeholders are aware and there are no future surprises.

The probabilistic framework makes sense but knowing what number is 50% achievable vs. 70% achievable seems hard. Any advice here?

• At Bessemer Venture Partners, we have 200 portfolio companies, so we can measure this. Of all our companies, how many achieved their revenue plan last year? For a single company, you can ask, how many times did we achieve our revenue plan in the last five years? Or, how many times did we achieve our forecast in the last five quarters?

Related to the question above, projecting revenue is usually much harder than projecting costs. Do you have any tips for teams who want to accurately project revenue?

• It depends on the type of business you’re projecting revenue for. It’s harder to project revenue for a consumer business than it is for a B2B enterprise software business. For a B2B enterprise software business, there is a relatively standard approach that involves looking at marketing leads, pipeline, quota attainment, and churn (including negative churn). Tools like Clari also exist to help Sales and Revenue leaders more accurately predict revenue.

• Most companies project revenue via a robust bottoms-up methodology vs. a simpler top-down approach. Ultimately, it can be useful to do both, but I would encourage doing the more diligent bottoms-up method at a minimum.

What about scenario planning? Does that come into play?

• Some companies routinely create multiple scenarios, while others have one approved plan with monthly updated forecasts. The best scenarios are built around specific events that are unknown now but will be known in the future. For instance, what if a large customer doesn’t renew? What if our new marketing campaign isn’t as effective as our last one? What if customers love our new product?

Different departments are often required to build a proper budget. How do the best companies manage this process across an entire company?

• At larger companies, it is the job of the Financial Planning & Analysis (FP&A) team, which reports to the CFO, to manage this process. As companies scale in size, the size of this team typically scales as well (though not necessarily linearly). For smaller companies that do not have a dedicated team, it is still the job of the finance function to create the budget. In those cases, in place of the FP&A team, the finance function can create a structure for other teams to work within. This may be a questionnaire or a spreadsheet with blank values that other departments need to fill in to help create the budget. In these cases, the finance team leads this process and project manages it as well.

Are there tools that can help with this process? When does it make sense to use them?

• We could write a whole post on this topic alone. Many tools allow companies - large and small - to create a proper budget. At the enterprise end of the scale, you have tools like Adaptive Insights and Anaplan. Other popular tools are Mosaic, Causal, and Runway. These tools allow you to add the numbers in, but knowing what numbers to add in is the real challenge - and where the Goldilocks framework hopes to be helpful.

What are the common failure modes for teams building budgets?

• One is that companies fail to get full stakeholder alignment on how the budget has been constructed and its probability of being achieved. Another is what I call Q4 Optimism Bias, which is based on the human tendency to believe that investments earlier in the year will begin to pay off by Q4. One more failure mode: operating teams think of the budget as something created by Finance, not by them. It’s important for all operating leaders to understand, and believe in, the company’s revenue, cost, and profit goals for the next 3+ years. Then they can see how their department fits in. If the sales leader asks for more budget, where do the dollars come from? From less R&D? From lower profits?

To finish, I want to ask you a question that I will be asking each of my interviewees, and that is what is your Steelhead goal in life? I.e. what’s really hard, takes a lot of time, but you know will be worth the effort?

• I’d love to have 12 grandchildren. I’m halfway there!

Thanks so much for your time today Jeff. This was great.