The Tech Sales Newsletter #76: SaaS spending patterns in 2024
After the great workload optimization of ’22-’23, last year was a bit of a mixed bag. On one hand, it went better than expected for most companies. On the other hand, growth definitely did not recover to pre-recession figures, and there is very little confidence across the tech sales community that this is likely to happen anytime soon, except for select sectors (i.e., cloud infrastructure software).
This overview is based on the 2025 SaaS Management Index Report by Zylo (thanks to Matt Harney from Cloud Ratings for mentioning it).
The key takeaway
For tech sales: Compensation plans are downstream from executive incentives, not market trends. In a landscape where single-year contracts are becoming the norm, any company tying a significant portion of your OTE to multi-year plans is likely prioritizing their financial optics over fair pay.
For investors: Future billings are no longer a reliable proxy for product stickiness. To identify companies with true staying power, a deep evaluation of their product-market fit and go-to-market strategy is essential.
It’s so over until we are back, baby
Source: Zylo 2025 SaaS Management Index Report
What a loaded chart, to say the least.
There are several trends we should address:
After 3 years of reduction in average spend per company, we finally have an UP year.
We are still far off from what companies used to spend at the peak of the tech bull market.
There are big differences in dynamics between SMB, Mid-Market, Enterprise, and “very large” organizations. Growth among the “whales” has actually continued over the last 3 years, partly driven by what the cloud hyperscalers have reported as a return of the focus on onboarding on-prem workloads to the cloud. For most “growth” companies in the 500 to 10,000 employee range, they actually reached their peak spend in ‘22 and cut a lot over the next 2 years, with few new workloads being onboarded.
Source: Zylo 2025 SaaS Management Index Report
Interestingly enough, tool consolidation played out very differently from market to market. The “whales” show very little sign of it (which probably is not surprising if we account for multi-year contracts), while across all other segments, there has been either partial or significant reduction in the use of different services. Some of those would’ve churned completely, while for others, the outcomes would’ve transitioned to another tool (i.e., consolidated platform usage).
Source: Zylo 2025 SaaS Management Index Report
On average, we should still recognize that companies use a lot of SaaS subscriptions (275 on average), and they still spend a lot of money on them ($49M on average).
Source: Zylo 2025 SaaS Management Index Report
One of the strongest growth dynamics is spend per employee—an area where we experienced 22% growth YoY in 2024. While headcount has remained static or experienced low growth at many companies, their investment in productivity improvements has continued. This is reflected in the area of spend most often associated with improved productivity: software.
When it comes to business models, seat-based subscriptions now account for less than half—a scenario that was unthinkable just a few years ago. As I’ve written about this previously:
My point of view comes from cloud infrastructure software, where the majority of vendors offer usage-based pricing models under a consumption (committed spend) contract. I believe that it’s the right model for those organizations since it allows them to appropriately scale customers with the right use cases. For more traditional types of SaaS products or services, there is now a growing discussion around outcome-based pricing.
One of the best examples is FinAI, the customer support agent by Intercom. This product is sold on the basis of case resolution rather than cases handled, meaning their clients only pay if the support ticket is closed successfully by the end customer. The initial case resolution rate of FinAI was in the 20% range, but over time it has improved; for certain types of end customers, it reaches a 90% rate. On an outcome-based model, the cost of support with Intercom is clearly increasing for their customers. However, the product is delivering the actual outcome they purchased it for, and it is evidently cheaper than achieving the same outcome with employees.
My personal view is that many “simple” SaaS products not tied to cloud infrastructure will struggle to compete against tools like ChatGPT, which can create an app. They’ll need an edge—essentially, outcome-based pricing. Very few will survive this change because, frankly, most don’t deliver the outcomes they claim they can.
As companies introduce a variety of consumption-style metrics, things can get…creative. Arguably, this can be beneficial for the reps since it makes it more difficult to do “apples-to-apples” comparisons.
The biggest victim of the tech recession is multi-year contracts. While they are still considered a staple of most compensation plans (CrowdStrike famously made it impossible to reach OTE by selling one-year deals), the reality is that most customers are hedging their bets and pushing for shorter commitments.
This shift is also directly tied to which business models can operate better under these conditions. In a consumption-based sale with a long-term roadmap of onboarding new use cases, it’s unlikely you’ll be suddenly displaced, even if it’s a one-year contract. The main impact of this shift in dynamics is on publicly traded companies, where CFOs often place an overweight value on future bookings. This, in turn, gets incentivized in compensation plans and drives rep behavior that contradicts where the market is heading.
Then, everyone is very surprised by the low attainment in otherwise high-growth tech companies (cough, Datadog, cough).
Probably the most interesting analysis in this report is related to who is actually buying the software in most companies:
IT is now responsible for just 26.1% of SaaS spend and 15.9% of apps. The remainder of software is brought into the organization by employees—often without IT’s knowledge or approval. This practice, known as shadow IT, now accounts for 3.8% of SaaS spend and 33.6% of apps. This is an 11.1% increase in spend and a 3.6% decrease in terms of number of applications.
Shadow IT spend is outpacing growth in the number of applications, which indicates that employees are buying fewer but more expensive apps. They’re likely opting for premium tools (often of the AI variety, as we’ve explored) or multi-functional apps with consolidated functionalities.
Because lines of business and employees purchase SaaS, IT no longer has the full picture of the organization’s software spending. As a result, SaaS spending grows unhindered, making it difficult to holistically and effectively reduce costs and risks.
Decentralized purchasing requires effective governance. Without clear policies and oversight, it easily leads to shadow IT, increasing financial, security, and operational risks. Gartner’s research highlights that many IT leaders struggle to implement governance in this environment, with 44% citing it as a key blocker to successful digital initiatives.
Now, what happens when a lot of these business users decide to just… use AI for text-to-application? Structurally, the demand and the will to circumvent “corporate controls” are present in organizations of all sizes. So, where would such optimization most likely start?
For example, HubSpot offers a calendar option, which means some customers may not renew a dedicated calendar app like Calendly. This approach allows companies to drive greater business value without the need to invest in a completely new tool.
Our customers’ rationalization efforts are an indicator of companies removing apps that don’t serve their needs and leaning in with vendors that are driving business value. And with increased overall spend, it’s likely these businesses are investing in AI or additional, premium features.
Since the “Great Rationalization” began in 2022, our data shows that the focus on reducing redundancy is paying off. For example, consider the top redundant function we commonly see in SaaS portfolios: online training classes. In 2022, online training classes averaged 18.5 applications. After rationalization, we saw an 18.9% decrease in applications with that functionality in 2023. Now, in 2024, the average number of applications dropped by another 5.3% to 14.2 apps.
How many “training platforms” will exist independently in five years, when an AI tutor on any topic is available to every business user through their internal “copilot app”? I wouldn’t be surprised if procurement teams make it mandatory for vendors to provide a training vector dataset along with a set of documents, which would then be ingested into these apps as part of the onboarding process.
The trade-off of the buying process moving toward business users is that the risks of investing in a poorly vetted product are significant. This creates a negative loop, where the first major breach caused by an application procured “on the side” will make the buying process for all future vendors more difficult. On the other hand, cybersecurity products that secure applications, data, or cloud infrastructure will continue to become proportionally more important.
We’ll close this article with one of the key recommendations from this report for companies:
If you’re unsure how AI is being used in your company, you’re already behind. Discovery is the starting point for managing opportunities for innovation and productivity gains while balancing potential security and financial risks.
IT and SAM leaders need to understand that AI will fundamentally change how SaaS is managed over time. While some traditional SaaS companies have already added AI products into their mix, more will continue to follow suit. Pricing models will change drastically, with many vendors shifting from per-seat pricing to consumption-based models for AI features.
Tracking how AI is consumed will be critical to ensure you stay within forecasts and budgets. Different pricing models and new complexities will make it challenging to monitor and manage AI, requiring new tools and strategies to predict costs.
Organizations should evaluate the value they receive relative to what their vendor provides. Without understanding how AI is being used and tracking the business value it is driving, forecasting overage charges will be difficult, putting your software budget at risk.
Effective governance is also essential to ensure AI is used responsibly and delivers value. Of the IT leaders we surveyed, 81.8% reported having documented policies specifically governing the use of AI tools. These policies may address guidelines on data security, permitted use cases, and other risk mitigation strategies.
But having an AI policy is just the start. To be truly effective, AI governance must be embedded in the organizational culture. This means educating employees on the risks and responsible use of AI, vetting all AI applications before purchase, and diligently evaluating AI features with vendors.
Interesting.
