Discover more from SaaSletter
SaaSletter - Morgan Stanley On The ROI of AI
Plus new Cloud Radio episodes + RepVue September data
Cloud Radio Podcast Episode
Maca is “the first operating system for value-based pricing” → value-based pricing is directly linked with software ROI - the theme for this article.
The ROI of AI Context
Anecdotally, the emergence of Generative AI has helped spark a rekindled interest in quantifying the ROI of software.
While other trends have increased the interest in software ROI, like:
ROI for New Deals: SaaS buyer surveys from G2 and Sapphire each documented vendor demand for ROI support while scrutinizing new app purchases
ROI for Renewals: On our podcast, Cory Wheeler (Co-Founder + CCO of Zylo) noted buyers beginning to request ROI proof at renewals
… the ROI of Generative AI discussion is more intensely quantitative.
Software is a wildly transparent and quantified industry. No matter your ARR level, a set of Robust benchmarks will exist: NRR, Magic Numbers, logo churn, quota attainment, sales per AE, and so on.
However, the universe of software ROI benchmarks is remarkably limited.
Cloud Ratings’ meta-analysis of 200+ 3rd party ROI studies is one of the only available:
“The ROI of Software + IT” - median 3-year ROI of 278%, payback of 6 months
“Drivers of Software ROI” - labor productivity is the #1 driver of customer value
The lack of ROI benchmarks makes sense:
ROI is very product and company-specific
ROI varies by customer (i.e. a larger customer gets a much bigger benefit from a fixed price suite)
A lack of metrics standardization
For mature categories, the business case is assumed to be proven with no need for doing the (not easy!) ROI homework
Generative AI requires ROI *math* due to:
Pricing: The rollout of AI features is creating price changes and challenging existing pricing models (see WSJ AI cost article below)
Skepticism Of New Technology: buyers want vendors to “show me your math” on AI productivity gains
Subscribe to SaaSletter:
Morgan Stanley On AI ROI
Quantified ROI is a key theme throughout Morgan Stanley’s excellent “AI Index – Mapping the $4 Trillion Enterprise Impact” (149 pages, released 10/1/2023).
Some excerpts oriented towards ROI:
Take rates - a metrics “cousin” of ROI - makes page 1:
FYI - those adoption cases are based on precedent 3-year adoption curves: 2% for public cloud, 18% for Office 365, 28% for ServiceNow Pro
ROI is literally #1 for their enterprise adoption drivers, with “confidence” and “visibility” the operative words:
Examples to date of ROI:
Value capture benchmarks - interestingly, the 0.3% - 1.0% value capture rate (equivalent to buyers enjoying 100x-300x annual ROIs) for GitHub CoPilot lines up with the WSJ note on its negative product gross margins → it seems very underpriced to lose money at those ROIs
Plus a sample of the granular math. Even in the initiation of coverage notes for “regular” non-AI software, this type + depth of ROI analysis is quite rare:
Morgan Stanley’s software TAM estimates by category. Notably, vertical software is the #2 beneficiary.
I enjoyed’s “The playbook makes me nervous” on the fact many SaaS companies are run in a “paint by numbers” mentality
Reflagging my July post - “AI In Finance + Enterprise” - that had an AI ROI section
ICYMI - we aggregated SaaS benchmarks from 2009 - 2021 to create longitudinal context: Historical SaaS Benchmarks
Plus our cloud cost optimization episode with Eldar Tuvey, Founder + CEO of Vertice: