Last week I wrote about how startups should very rapidly alter their desire to be unicorns and instead aspire to be a cockroach. Survive first to have a chance of thriving down the road.
Coincidentally, that same week
wrote an article that covered the same topic.The main difference - which I agree with - between my article and this one is the need to balance growth with efficiency.
While the growth-at-all-costs mindset is dead, growth is still really important if a company plans to fundraise again. Investors just now want strong growth AND efficiency. If a company isn’t ”default investable” based on its metrics then they need to do one of the following:
Extend cash runway long enough to get to “default investable”.
Move quickly to becoming “default alive” before running out of cash
My original intention for this post was to address the manners in which a head of engineering can contribute to the “default alive” scenario. This would be mostly concerned with reducing cash burn. There is another angle to this topic, though, and that’s growth.
Whilst the current environment calls for austerity and cash preservation, that alone is insufficient. Startups will need that AND growth. Both of these variables are ones that heads of engineering have a substantial influence on. My post today will focus on the austerity measures with next week’s post focused on growth.
Optimization targets for cost cutting
Cutting costs is a balancing act. Cut too much and you risk crippling the business. Conversely, cut too little and you also risk harming the business. When cutting costs I keep a few targets in mind. The first is R&D spend relative to the size and stage of the business. The second is margins, which is very relevant for SaaS companies.
The data for R&D spend relative to revenue (and stage) is available, albeit from public companies. This data can serve as a proxy for late stage companies.
R&D spend is 24% of revenue. 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) For many of these companies, they were in their Series B or Series C two years prior to IPO, so it’s safe to say that spending a quarter of revenue on R&D is the right level for a SaaS business even at earlier stages. No matter where your SaaS business is in its lifecycle, as one founder put it to me, “managing a large and growing stack for a cloud application is damn tough” so you’re going to be spending materially on the stack no matter how fast you’re growing or how mature you are.
Source: Sammy Abdullah
The second target is margins, gross and operating. Gross margins for most software companies will be the raw materials of your products and the ongoing cost of support. Raw materials are broadly speaking hardware costs - if your product has hardware components - and infrastructure for SaaS companies. How well you build and manage your products will influence your gross margins. On the other hand, operating margins will mostly be driven by labor costs especially for software engineering organizations. Margins aside, there is another key financial metric you want to target: cash runway, which is influenced by sales and margins.
The target margins and cash runway you are aiming for will be stage dependent and defined by your CEO/CFO. Defining them give you and the rest of the company concrete targets to aim for and provide a good context for cost optimization.
What to cut then?
With some key financial targets in mind, now comes the time to make the cuts that will facilitate achieving these outcomes. When cutting costs I use the following 2x2 (of course!), which helps prioritize which areas to reduce spend in and why. Some areas yield the most cost savings, but are harder than others and vice versa. Quadrants market in green are where I start first, yellow next and red I try to avoid.
Low pain + high impact
The most common cost savings that fall under this category are infrastructure costs, be those that power your dev/test environments or your production SaaS. I am willing to wager that most software companies can find 30% infrastructure cost-savings. Simple things like cleaning up cloud resources after use, stopping unused ones overnight, and right-sizing resources will go a long way in reducing costs. Revising some of the scaling heuristics is also beneficial too. Perhaps your auto-scaling groups are underutilized, or your capacity/health checks thresholds are wasteful.
AWS Cost Explorer, or its equivalent on other clouds, should be part of your daily routine. I have found this tool to be immensely useful. All the data you need to help optimize cloud costs is available through these tools. You should put them to use.
Another area that is ripe for optimization is SaaS proliferation. You must rationalize and consolidate your SaaS usage. Trello, Jira, Basecamp, Notion, AirTable and Asana aren’t that different. You don’t need them all. Pick one. The amount of waste here can be substantial.
The savings you find here will impact both gross and operating margins.
High pain + high impact
There’s no way to avoid cost savings without bringing up RIFs. The largest costs in an engineering - and startup - budget is headcount. This option will yield the most savings, but is obviously the hardest to do. I won’t go into details on how to do a RIF, both in terms of criteria and process. One important factor when thinking about a RIF is it’s impact on growth. There’s an obvious and immediate impact of a RIF on cost savings. The impact on growth is not obvious at the outset. Cut too deep and you risk hampering the future growth of the business, which in engineering terms is the ability to ship more products and features. You need to balance the need to cut costs with growth when conducting a RIF. Unlike auto-scaling groups, headcount can’t scale 30% overnight 🙂
Labor costs, or more generally speaking a RIF, impact operating margins. Some groups like support will also impact gross margins.
Low impact + low pain
These are the miscellaneous cost items in your budget. They include patents, T&E and the like. Depending on the magnitude of your spend here, the impact will be mostly low both in terms of pain and cost savings. For some other groups, like sales, the impact on costs savings from T&E can be substantial.
There could very well be other costs savings that are idiosyncratic to your organization. You should make sure that you explore all the major driers and optimize accordingly. Generally speaking my experience has shown that infrastructure, tools and headcount are contribute upwards of 90% of the costs incurred in a software engineering organization. With headcount being by far the largest contributor.
I can’t stress how critical this next phase is. There will be a reckoning in startup-land sometime later this year. Many of the late stage companies who raised capital in 2021 - at peak valuation - will be attempting to raise capital later this year. Many won’t be able to and will die.
The funny thing is I had been working on the article for a couple of weeks and I came across your blog accidentally. Love your newsletter. Big fan!
Karim - Great piece. Timely, insightful + actionable. Will be sharing this a bunch. Thanks.