How CFOs think about software spend
For the love of God, please stop it with the Miro licenses
Today I want to talk about software and tooling spend from a non-technical perspective. But first, I’ve got to level with you - I don’t know what half the things we buy “do” (Stitch vs Fivetran anyone?)
What I can attest to is that there’s a business case behind why we need each tool, supported by an executive who is directly accountable for results. We trust leaders to make decisions on what tools they need to get the most out of their teams.
At the end of the day, software is about enabling people to build. And you can’t separate the people from the technology. If you change the people, you change the technology, and vice versa. The improvements come from improving one or both. But changing one or both will change the other as well. (Paraphrasing a knowledge bomb this product guru once dropped on me…)
But back to budgeting…I look at Finance, and FP&A specifically, not as software decision makers (in most cases), but as the brokers between the business unit, the company P&L, and the vendor trying to gain our biz.
It’s the finance department’s job to understand:
What type of tool it is - Critical infrastructure, productivity, communication, security, G&A
How we pay for it - Usage based, headcount based, license based, transaction based, hybrid
When we pay for it - Annually up-front, monthly in-arrears, quarterly up-front etc.
Unintended consequences - Do we have the bandwidth to implement this? Are there any other tools this takes the place of?
Types of Tools
I’ve taken a shot at bucketing these in relative order of importance. When someone brings a tool to me and asks for budget, I instantly think - Is this a vitamin or a painkiller? In other words, does our business cease to exist without this? Does this improve something we already are doing? Is this a nice to have?
The following is a general stack ranking of tools by “bucket”.
Tier 1: Keep the Lights On
Critical Infrastructure (AWS, Snowflake, GCP)
Security (Crowdstrike, Jfrog, DataDog)
Engineering (Gitlab, Docker, Postman)
Finance / Accounting (Quickbooks, NetSuite, Xero)
Money In, Money Out (ADP, Bill.com, Ramp)
Communication (Slack, Gsuite, Zoom)
Tier 2: Run the Business
Workforce Management (Workday, Bamboo HR, Gusto)
Sales (Salesforce, Docusign)
Customer Support (Ask Nicely, Zen Desk, Churnzero)
Marketing (Hubspot, Mailchimp, Marketo)
Finance (Carta, NetSuite, Ramp)
Recruiting (Greenhouse, Workable, Linkedin)
Databases (Tableau, Grafana, MongoDB)
Design (Figma, Fullstory, Adobe)
Tier 3: Run the Business Better
Project Management (Miro, Clickup, Asana)
Budgeting (DataRails, Adaptive Insights, Vena)
Expense Management (Expensify, Tripactions, TravelPerk)
Meeting Management (Calendly, Chilipiper, Vidyard)
Analytics and Nice Charts (Sigma, Looker, Chartio)
Sales Efficiency (Gong, Zoominfo, Linkedin Sales Navigator)
Research (Pitchbook, Crunchbase, CapIQ)
People Tools (Lattice, Culture Amp, WorkHuman)
Modern Marketing Tools (Jasper.ai, SproutSocial, Podium)
Treasury and Procurement (Kyriba, Coupa, Ivalua)
How do we pay for it?
In the words of Mac Miller - “CTC…Cut the check!”
This informs how we budget for it.
Generally speaking, the easiest tools to budget for are those linked to headcount. What can be tricky, though, is nailing down if the tool is exclusive to one department. For example, if everyone needs a Gitlab license in engineering, that’s a pretty finite group to forecast for. You just scale the spend with the department’s forecasted headcount each month.
But with cross-team collaboration tools like Miro, this can go haywire real fast. For example, someone in Biz Dev just sent me a real neato market landscape overview they spun up in Miro. I clicked to view their book report, only to find out later from the head of Product (who “owns” the tool) that I (a member of finance) am now chewing up one of our paid Miro licenses. License proliferation is a real thing.
But, I digress. Here are the three most common ways to pay for a tool:
Seat Based Subscription (Recurring): The tried and true. You pay, say, $8 bucks for each person who uses Slack per month.
Consumption (Usage Based): Now this can have different flavors. A lot of times you pre-pay for credits to get a better rate. Other times you just get billed based on a flat rate at the end of each month. AWS and Snowflake are the easiest examples.
Hybrid (Mix and Match): ZoomInfo is the king of this. They blend the two models above. They charge you a per seat subscription fee and then layer on a fee based on credits. It’s genius (for them).
Transaction Based (Take Rate): You pay a percent of whatever your transaction volume is. Stripe is the most common example. You get hit with ~3% per transaction. Quickbooks can also charge you a variable fee based on stuff you bill (in addition to the subscription).
Big Tony’s Enterprise All You Can Eat Tier: The company gets unlimited access to the same shared resources and just pays one big fee. This is typically reserved for, say, the Walmarts of the world who can buy tens of thousands of licenses. You gotta spend a lot to get this type of deal, and usually commit for multiple years.
When do we pay for it?
When I go into a negotiation, my primary lever is trying to knock them down on price. But when that well is tapped, the next lever I go to is payment terms.
Even if we are accruing the expense each month of the year from an accounting standpoint, I also want to look out for my balance sheet and try to give it some breathing room.
Now, these two levers are not isolated. Sometimes I will pay for something annually, up-front, if I can get better pricing. It’s a tradeoff you should be willing to make if they come down enough.
It does eventually become more art than science though as to where the approach shifts.
When you think about the length of the contract, consider if your organization is forecasted to grow and if you will be able to return to the negotiating table in a year to get better unit economics.
What you have to weigh with that consideration is the impact of inflation. Sometimes it’s better to lock in a long term deal because a.) the discounting is better and b.) you think they’ll raise prices due to macro conditions.
Unintended Consequences
Don’t Double Pay
What are we currently using today?
What goes away?
Does this replace any existing tools?
Is this duplicative?
You don’t need both Asana and Monday. Pick one.
Users impacted
Will any other groups use this tool?
Have you talked to them?
Spoiler alert: The answer is usually “Um, no(t yet)”
Don’t sign another leader up for a tool they didn’t demo. And similarly, don’t cancel a tool they are using because you want a different brand that does the same thing.
Hidden Fees
Are there any one-time platform fees?
Are there any implementation fees?
Do we have to hire a third party to install this stuff?
Is this “enough” capacity we are signing up for? Will we blow through this initial tier? Have you back tested it with a forecast?
It sucks when you have to pay a huge implementation fee. Once at a company I worked at we hired a third party to implement NetSuite for us, and it was 2.5x the cost of the first year subscription. Yikes.
And I use the phrase “implement for us” very loosely. The implementation became 30% of everyone’s full time job. We lost time and money in order to spend money on a tool to mange our money. And imagine ripping that thing out after FINALLY getting it installed? You’ll replace the office carpet twice before NetSuite goes anywhere.
And sometimes costs are hidden in the sense that you are signing up for an artificially low tier. Don’t get suckered in - if it feels too good to be true, it probably is. “It’s so cheap! It only starts at $1,000 per month for 100 documents!”.
OK, but we send out 200 per month, Dale.
And by then, they’ll already have their talons in and be riding your growth to the promised land of quota achievement.
Bandwidth
Do people have time to get value from this tool?
Is it just for people managers?
Do people have time to implement this tool?
Are there any other implementations going on this quarter?
You don’t want to buy Gong if you have a flat sales hierarchy and there’s no one with the time to listen to the calls.
And in my experience, you don’t want the org to implement more than one major tool simultaneously. All the project managers will be booked up and people are busy enough doing their day jobs. Plus, you don’t want to move two pieces of cheese at once. Especially if they depend on each other. One time we tried to knock out NetSuite and an FP&A tool simultaneously. As you’d expect, it was like trying to nail Jello to a tree.
Fine Print
Is there an auto renewal clause?
Do they get to use our logo in advertising materials?
Read the whole contract. I know, it’s boring. But you can’t really outsource this part. The buck stops with you (well, unless you hire McKinsey or Deloitte to make the decision, then you have culpable deniability).
What I’ve been reading
FP&A gave me a super power - it allowed me to ask anyone in the company anything about their department. When you broker the budget, you can ask the questions. It’s like a license to learn. And about 17,793 questions later, this got me to CFO. Needless to say, I think FP&A expedites your learning curve.
My friend Nicolas Boucher is sharing awesome FP&A tips on LinkedIn. In particular, I love his infographics which make complex concepts easy to read and absorb.
Nicolas has helped more than 200 people get into FP&A with his online video course.
If you want to move into FP&A, this is what you need!
Quote I’ve been pondering
“A man who makes the trip to the Caribbean by land once is a hero, the man who does it twice is a fool”
-The Fish that Ate the Whale, by Rich Cohen
Mostly archives
Building Finance Teams: The lifecycle of beancounters
The Game Inside the Game: PE vs VC
What Product Market Fit Feels Like: More than a number more like a feeling
Amazing post CJ! Super helpful for founders to understand how finance influences companies buying products!
This is great. The part I'm really curious about, maybe for another post, is how the discussion with the leader wanting to buy a tool goes, whether they really need to have a business case, and how much you poke at it. Do you end up having conversations like whether you really need Tableau or Miro, and how does "really need" end up getting defined? Do you diplomatically call BS if someone says the company growth rate will be 1% higher if you buy Podium (just picking an example out of the hat)?
In my experience, leaders often have pretty optimistic business cases. Many times, they're also difficult to define or measure. How can we tell if we're collaborating better with Miro?
Anyway, I know this topic could fill a book, but I'm mostly curious about the war stories.