Sailing the 7 seas: Tracking and benchmarking R&D budgets over time
Sailing the 7 seas: Tracking and benchmarking R&D budgets over time
Part three of the trilogy
My previous two posts addressed top and bottoms up budgeting, the two predominant methods used to derive budgets and operating plans.
“Everyone has a plan, until they get punched in the mouth” - Mike Tyson
However, a budget isn’t just a one-time exercise that occurs towards the beginning of a fiscal year. A budget is a plan, and as we all know, plans can change. Hence, budgets, or more specifically actuals, need to be periodically monitored and adjusted to ensure that the company is headed in the right direction. We’ll be walking through this process in today’s last post on this topic.
Benchmarking your spend
Before we dive into the budget-to-actuals process, I’ll go over a final step I like to conduct before finalizing a budget: benchmarking.
Once you have a finalized budget, or operating plan, for your company, you should be able to express spend across various categories (e.g. R&D, CapEx, S&M..etc) in terms of % of sales. This metric is very helpful as it allows you to evaluate your spend relative to comparables. Said otherwise, you can look at your spend, expressed as a percentage of sales, and compare it to companies similar to yours. This is useful to judge, and potentially justify, your spend. I use this as a tool and not necessarily as a guide towards the level of spend I commit to in a fiscal year. There will be many factors that can influence the level of spend. Perhaps your company is in a very high growth stage and needs to increase R&D spend beyond normal levels. Again, use this data to stress-test your spend, but not to derive it.
There’s an abundance of data, from both the public and private markets on the level of R&D spend, both across startup stages (ARR) and over time (pre/post IPO). Two data points I like looking at are this post from Sammy Abdullah, which shows that SaaS companies spend about 24% of their revenue on R&D.
Leading up to the IPO, SaaS companies spent on median 24% of revenue on R&D. As you can see there is almost no deviation between the financials reported at IPO and 2 years prior (medians were 23% and 23% respectively) - Source: Blossom Street Ventures
This data above is biased towards large, late stage software companies. Luckily we have extra data - via ICONIQ - that looks at R&D spend across early stage software startups. The data shows that R&D spend is significantly higher the earlier in the startups’ life-cycle.
One other area I like to benchmark is not just aggregate spend, but categories of this spend. For most R&D organizations the bulk of the spend will be People and Infrastructure. Again, we have ample data to help with this benchmarking exercise.
Budget-to-actuals
A budget-to-actual is a periodic process, typically done once a quarter, in which actual spend is evaluated against the planned budget. I typically conduct one of those every quarter with my Finance team. The intent is to look for big variances (Variance = Actual/Budgeted - 1) between the plan and actual spend.
It’s important to note that any large variance, be it above or below budget, is critical to understand. Dramatically overspend and you risk missing on your fiscal goals. Similarly, dramatically underspend might also hinder your ability to deliver on your plans, especially if the miss is due to not being able to hire.
The illustrative data below depicts the budget-to-actuals of an R&D organization. The spend is tracked across the major categories, which in R&D tends to be People & Infrastructure. Variance is tracked both in aggregate and per category to signal any large deviations. Cells in red indicate a variance >= 5%.
One important note. It’s very helpful to look at this data on a monthly basis, but it must be complemented with an aggregate year-to-date view as well. Consider the case of January’s People spend in the data above. January’s spend was ~6% below budget, presumably due to missing hiring targets or timing of new hires. February, or some other future month, might have a higher variance due to compensating for the January miss. This means that some future month might have a large variance in the People category, but the overall spend still remains in check. Another common variance I see is in infrastructure spend. You might purchase a portion of your cloud spend ahead of time (commitment) which might result in a large variance in the month this spend occurred in, and hopefully lower spend down the road. You need both a snapshot in time view of your spend and the aggregate view to be able to fully assess variances.
Finally, the regular budget-to-actual process is also very useful if spend needs to revised upwards or downwards. For example, if a company is hitting its targets and feeling more optimistic about it’s future growth it might decide to increase its spend in some areas. Conversely if things aren’t going according to plan, budgets will be trimmed and the actual-to-budget process can highlight areas to trim.