How to manage runaway SaaS spend

70% of costs at SaaS companies walk on two legs.

Take the P&L from any venture backed tech company, and you’ll realize that anywhere from 2/3 (on the low end) to 3/4 (on the high end) of all spend is wrapped up in what I call “total effective labor.” That’s the aggregation of payroll, commissions, benefits, bonuses, contractors, and consultants your company hires … the “people.”

That leaves roughly ~30% of company spend that’s wrapped up in “stuff” without a heartbeat.

Each department within the org has a different mix of non-people “stuff.” It’s common for teams to call people spend “direct” spend, since it’s linked directly to people, and broadly classify everything else as “indirect” spend (kinda simple, I know).

But we can go a level further and come up with some general rules of thumb, or ratios, at the department level to split out those direct and indirect costs.

Cost of Goods Sold:

• 33% people

  • Customer support and Customer Success (if not in Sales)

• 33% tools or data

  • Customer support tools (Ask Nicely, Zendesk, Churnzero)

• 33% hosting infrastructure

  • AWS, GCP, Snowflake

Sales:

• 75% people

  • Payroll + commissions

    • Note: If you have commissions showing up in any other department, you are doing it wrong. Commissions should be reserved specifically for sales. Any other variable comp should be classified as a management by objective (MBO) bonus.

• 15% tools

  • RM (Salesforce, Hubspot)

  • Sales efficiency (Gong, ZoomInfo, LinkedIn Sales Navigator)

  • Meeting management (Calendly, Chili Piper, Vidyard)

  • Contract management (DocuSign, LinkSquares)

• 10% travel

  • Revenue-generating travel: Client-facing meetings, conferences

  • Non-revenue-generating travel: Small hands, all hands, QBRs (quarterly business reviews)

Marketing:

• 50% people

• 40% programs

  • Includes tools, advertising, campaigns, swag, conference sponsorships

    • This varies by company scale — you need people to deploy the marketing programs and campaigns, so at first this skews towards people (60% / 40%).

    • But over time, the ratio starts to slide in the direction of programs (40% / 60%) once you have people to spend it.

    • Generally speaking though, you won’t see it move more than 70% / 30% in either direction.

    • Tools you’ll commonly see pop up in your vendor report include Hubspot, Mailchimp, Marketo, Jasper.ai, SproutSocial, Podium.

• 10% travel (and shipping)

  • Lead-generating travel: Conferences, and anywhere you ship that booth

    • Tangent: Sadly, it’s sometimes cheaper to buy a 42-inch Vizio TV for the booth at a local Best Buy once you land, and then just leave it in the hotel room as a gift for the hotel staff, rather than paying to ship it all the way back.

    • A CMO once told me he had been doing this all year. I got wicked mad at the waste, ran the numbers, and discovered he was right by a factor of almost 2x.

Product:

• 70% people

• 25% tools

  • Project management (Miro, ClickUp, Asana)

  • Design (Figma, Fullstory, Adobe)

• 5% travel

  • Customer studies: meeting with clients firsthand to understand their needs. This may include advisory committees.

  • Conferences: often working hand-in-hand with marketing to man the booth

• Internal travel: Product people really love their small hands meetings, where they talk about activation metrics and gather around campfires to play the banjo.

Engineering:

• 93% people

• 5% tools

  • GitLab, Docker, Postman

  • Surprising tangent: A lot of dev tools are actually relatively cheap (as a CFO, I rarely ever say this. It actually hurts that I’m writing this).

  • Most start as free. And if you are an org that’s under 200 people, there’s a good chance that more than half the tools your engineering team is using are still free.

• Very minimal travel (2%?)

  • If your engineering team is traveling a lot, something is very broken.

  • There are no customer-facing opportunities for engineering teams, and Product and Marketing people should be representing the company at conferences to better communicate the story (sorry if I’ve offended any Toastmaster developers)

Finance

• 60% people

• 25% tools

  • Money in, money out (ADP, Bill.com, Brex)

  • Finance/accounting (QuickBooks, NetSuite, Xero)

  • Treasury and procurement (Kyriba, Coupa, Ivalua)

  • Cap table management (Carta)

• 12% professional fees

  • Lawyers (Cooley, Fenwick, Goodwin, Wilson), Tax (BDO), Audit (Big Four)

• 3% travel

HR

• 80% people

• 10% recruiting “advertising” / “program” spend

  • LinkedIn, BuiltWith, Indeed, Monster, Glassdoor

  • I think of this as effectively marketing program spend but for the org’s overall brand to attract talent

• 7% tools

  • HRIS tools (Workday, Bamboo HR, Gusto)

• 3% travel

“G&A” / “Exec” / “Operations”

• 40% people

  • CEO, COO, admins, facility workers, etc.

• 30% tools

  • This is where I’m throwing shared tools (Zoom, Slack, GSuite) rather than allocating out to the departments

    • As you get past ~$25M in ARR you can start to allocate the larger shared costs (like rent) so G&A doesn’t look like an absolute albatross

  • And I’m also throwing all security and compliance tools here (CrowdStrike, Jfrog, Datadog)

• 20% rent and overhead:

  • Also includes utilities, office expenses, snacks, the disgusting Flavia coffee I had to drink at PwC, etc.

Nothing like some bagged coffee to get you through that 92-page RFP.

• 10% exec travel

  • This is where you park the 150-hour NetJets card for your CEO and hope you get invited someday when you beat your revenue forecast.

Taking our eye off the ball

Many of the tools move inline with the people — you can get a good grip on your Figma forecast by attaching it to the number of heads you think will be employed each quarter on the Product team. The same goes for Salesforce, Slack, Gong, and other tools that are easily attributable to real people in real departments.

But what I’ve noticed is that we spend more than 70% of our time on the 70% of costs wrapped up in headcount.

Why is that?
1. Headcount has the most “juice” to squeeze when you want to make a change.
2. Travel, software, and rent are tied to headcount, so you can’t inform one without the other.
3. You can cut people faster than you can bail out of annual software commitments
4. And most importantly — headcount can talk back and tell you when it wants more headcount on the team (Miro is slightly less vocal, but getting there).

But based on my research, it feels like there’s actually more juice than we might think in the SaaS bucket, as it’s growing even faster than headcount at most companies.

Source: Vendr

The average organization uses 110 distinct SaaS apps in 2022, which increased to 130 in 2023, per Vendr. And when you size up SaaS applications on a per employee basis, it’s jumped from 8 to 12 from 2015 to 2022.

And the chart above speaks for itself — not many companies are growing headcount by the equivalent of +60% y/y.

That’s why the top three things that keep me up at night are:
1. Stacking bodies
2. Orphaned and duplicative applications
3. Bears

Number #2 is increasingly becoming #1B when you factor in the rate of change applications are undergoing and the leap of faith we take on multi-year contracts. I constantly am asking myself:

“Will the application I’m signing up for still be best in class before the contract is up? Will my headcount forecast be right? Is the level of discounting I’m getting worth the cost certainty? Can bears really smell through three feet of ice?”

Most employees are at-will. Most SaaS applications are 1.7 years.
I think a lot of CFOs realize their exposure to SaaS apps, and the underlying lack of flexibility, when the water in the Amazon fell and the piranhas came out (read: when the economy took a nosedive and companies stopped expanding their seat counts).

Oh, the irony

It’s a bit ironic that the biggest culprits of SaaS waste are SaaS companies. If you sell SaaS, you also waste SaaS. The stats I’ve referenced are for SaaS companies that are SaaS users.

In fact, about one-third of SaaS software spend is underutilized or wasted, according to Flexera’s State of ITAM 2022 report. It’s kinda like the fix is on — SaaS companies sell to other SaaS companies in an inefficient way. We’re all complicit.

Once we admit that to ourselves, here are five steps to manage runaway SaaS spend:

1. Consolidate the 30% on one platform

• Brex's procurement capabilities keep much of the runaway SaaS vendor spend in check through advanced budget controls across invoices and purchase cards, without slowing down key business purchases with all the red tape of a purchase order.

• You might recall that Brex offers similar controls for travel and expense management, so you'd also be able to keep those T&E line items in check on the same platform. You basically get the superpower of "visibility" for much of your non-payroll spend. It might not be as cool as the power of say, invisibility, but you'd be surprised what today's finance leaders wish for.

2. Weed out reimbursement requests for single licenses

• As an FP&A professional, single licenses of Dropbox, GitHub, and Microsoft Office have dogged me. There’s a good chance that you are already paying at the enterprise level for licenses, and this person was just lazy and didn’t want to ask IT to provision a license.

• Or, even worse, your org doesn’t want this person using that tool at all and they are going around guardrails.

• I’ve personally been that guy when they tried to take my PPTX away from me in favor of Google Slides. I’d rather die.

3. Export vendor lists for department leaders on a monthly basis

• Every month I meet with the CEO’s direct reports. And ahead of that meeting we send them an export with every vendor name and the associated spend linked to their department.

• For example, this way the CPO can scan the list and see Miro, Dovetail, Fullstory, etc. and call out any new ones that have popped up on the list and been attributed to their department.

• You need eyes on the vendor names to appropriately manage the spend.

4. Offboard people who leave the company (maniacally)

• It happens more than you think — an employee leaves the company and eight months later you learn you’ve been paying for their PitchBook license the entire time (I’ve been both a victim as an FP&A leader, as well as a beneficiary as a former employee of this phenomenon).

• Setting auto reminders for employee offboarding is critical to make sure you aren’t paying for extra licenses. I get pinged every time someone leaves so I can kick them off our TravelPerk account. The same should happen for Salesforce.

5. Build in better approval workflows

• To weed out shadow IT, put the responsibility back on leaders.

• It’s funny how quickly people become your advocates for cost reduction when there’s a visible approval chain linked back to them.

• Tools like Brex can help you set up the proper workflows to not punish, but enable, your leaders to help with procurement.

This post originally appeared on mostlymetrics.com.