During a recent ROI calculation study for a B2B SaaS client, we looked at how their direct competitors had generated the ROI figures published on their websites. Nearly choked on our coffee when we read the fine print: sample sizes as small as 3 customers and typically around 5 customers. No information on selection criteria. What the heck is going on? Are our customers wasting their budgets on hiring us to actually perform credible ROI research?

SaaS ROI calculations are inescapably everywhere: on landing pages, in sales decks, investor updates, analyst briefings, procurement justifications. They’re meant to be clean, quantifiable, and hard to argue with.

But in most software categories, those figures are no longer trusted.

Savvy buyer executives have learned to discount ROI figures by default, even if they haven’t read the fine print about how the numbers are generated:

This is the current standard of practice.

It’s not that the numbers are made-up. It's that super-limited results under ideal conditions rarely survive real-world scale.

Post-sale, adoption rates drop, integration challenges surface. Teams revert to old workflows. The promised gains are delayed, diluted, or disappear.

In B2B, internal CRM and Martech tools routinely fall short. Salesforce’s own ecosystem shows this. Nearly two-thirds of implementations fail to meet their original goals, mostly due to poor adoption and misaligned processes — not the tool itself.

And in Edtech, even Randomized Controlled Trials (RCTs), often promoted as gold-standard proof, suffer from short durations, novelty effects, and ethical constraints in the real-world. Even when statistically significant student gains are shown, they often vanish when the tool rolls out more broadly.

Buyers know this and the result is quiet cynicism. While ROI claims still show up, few senior buyers take them seriously. They're scanned, but not fully believed. They make the deck look complete but rarely inform a real decision.

And when some buyers do believe them — when they use those numbers to justify a purchase — the disillusionment that follows hurts everyone: the buyer, the vendor, and the credibility of the entire category.

That’s where we are now.

ROI numbers are being generated from narrow inputs under ideal conditions. A handpicked team. A controlled rollout. A limited timeframe. The figures are technically true in that context, but they don’t survive outside it.

Other than ridiculously small sample sizes of course, three other patterns show up across failed or disappointing implementations.

Fragile AttributionBuyers don’t operate in lab conditions. When five initiatives launch at once — new software, process change, leadership turnover, new incentives — it becomes difficult to isolate cause and effect. Vendors often attribute all improvement to their product. Buyers sometimes let them, because they want a clear story. The result is an illusion of certainty.

Conditional ResultsEven the best products rely on other conditions being in place:

If a 20% productivity boost is achieved only with 100% adoption and full-feature usage, it’s unlikely to replicate at scale. When early pilots are used as the reference point for enterprise pricing or future savings projections, the bar is often set unrealistically high.

Missing CounterfactualsMany ROI stories skip the question of what would’ve happened otherwise. Maybe the team would have improved anyway. Maybe a cheaper tool would have delivered 90% of the benefit. Without counterfactual thinking, the ROI figure becomes inflated by default.

These weaknesses show up in retention metrics. In B2B SaaS, many churn conversations begin with: “We didn’t get the value we expected.”Not: “The product failed.”Not: “It lacked features.”Just: “The expected gains never materialized.”

In EdTech, school leaders sign multiyear deals on the back of glowing case studies. But when results don’t match the promise — especially for struggling students — districts quietly pull back. Teachers stop using the product. Renewal becomes a fight.

And in enterprise contexts, ROI fails when it’s treated as a promise rather than a hypothesis. When the story is too tight to leave room for real-world conditions.

The vendors with high LTV are the ones who build their credibility slowly.

Best practice is to stop treating ROI as a headline number and use it as a starting point.

Buyers who ‘get it’ ask different questions early in the process:

They push for detail, not just data. They want to know what would happen in their environment, with their constraints. Not in a pilot, not with a cherry-picked usage cohort.

When buyers set expectations this way, they create the conditions for better implementation, better adoption, and fewer surprises.

They also avoid the cycle of enthusiasm → disappointment → churn.

On the vendor side, some have started to treat ROI like a product. Not just something to declare, but something to deliver.

They do this by:

In B2B, some vendors are rethinking what qualifies as ROI. Instead of pointing to surface metrics like web traffic or booth scans, they’re building methods to track pipeline influence and revenue attribution over time.

One sales enablement platform, for example, worked with its top customers to trace which content assets were actually used in won deals. They then stopped reporting on “engagement” and started publishing reports showing time-to-close improvements and win-rate lifts by segment. Not every client saw the same result — but that’s what made the data more credible.

That kind of approach is powerful. When everyone else is selling generic gains, the vendors who show specific, verifiable impact get noticed. Done right, the very credibility of your ROI calculations cane become a differentiation opportunity, a competitive weapon, not just a checkbox.

Another good example comes from the few EdTech vendors who’ve moved away from “engagement metrics” and now publish longitudinal impact data. They show year-over-year changes in proficiency across student subgroups. Not always flattering — but more trustworthy.

And then there are vendors who lean into Return on Failure — tools that protect you from risk, not just generate upside. For example, a compliance vendor might not “increase revenue,” but it can show what happened when a client avoided a $20M class-action lawsuit due to better tracking.

That’s real money.

These vendors win trust because they respect the buyer’s intelligence and they build a case that holds up. And if the numbers ever fall short, the relationship doesn’t.

Because there was no bait, there’s no switch.

This shift is about being durable. Because in every SaaS market, buyers are under pressure to justify every line item. Vendors that get this, keep customers.

And it can be more than just about retention. For vendors that do have credible ROI stories, this is a chance to go on offense. If your competitors are still telling thin, unverified stories, they’re giving you a massive opening—if you have real evidence, (more) rigorous methods, and the confidence to show your work.

Adil Husain is the founder of The Intelligence Council and Managing Director of Emerging Strategy. He has over two decades of experience advising education and learning organizations within the K-12, Higher Education and Workforce/Professional space, including leading universities and 13 of the 20 largest U.S. education companies by revenue, on corporate strategy, product-market fit, customer acquisition, and growth.

You can contact Adil here directly, or connect with him on LinkedIn.

The Intelligence Council publishes sharp, judgment-forward intelligence for decision-makers in complex industries. Our weekly briefs, monthly deep dives, and quarterly sentiment indexes are built to help you grow your top-line and bottom-line, manage risk, and gain a competitive edge. No puff pieces. No b.s. Just the clearest signal in a noisy, complex world.

Read our ‘manifesto’ here.

We deliver must-read insights to strategy, product, and commercial leaders within education firms.