By Sean Simon, Co-Founder, Blurbs
For an industry built on measurement, martech has a way of dodging the most important metric of all: whether it is actually working.
There are currently more than 15,000 martech products available in a growing market that’s on track to top $215 billion by 2027. But McKinsey recently interviewed about 50 senior marketing leaders at Fortune 500 companies and found that not one could clearly articulate the ROI of their organization’s martech investment.
That study is not alone. The CMO Survey found that only 56% of purchased martech tools are actually being used. In other words, nearly half are sitting there, paid for but largely idle. Respondents rated martech’s impact on company performance at 4.7 out of 7, with actual payoffs coming in 34% below what leaders had hoped for.
More tools, more spending, more dashboards and still a basic question hanging in the air: Why does so much of this feel broken?
It stems from the torturous way buyers and vendors have to connect.
Buyers and vendors spend too much time performing procurement and not enough time having honest conversations.
Start with prospecting. AI has not improved the average email pitch. It has industrialized it. Even leading vendors still cling to outdated practices like batch-and-blast email campaigns and limited personalization. Now those same habits can be executed at an absurd scale. Meanwhile, the inbox itself is changing. New AI features can summarize threads, answer questions about inbox content and filter out unimportant emails so users see what matters most. That is good news for recipients and bad news for every vendor still spraying automated messages into the void.
That would be less of a problem if buyers were eager to talk early. Most are not.
According to 6sense’s 2025 B2B Buyer Experience Report, buyers now travel a long way down the path to purchase before they ever engage a seller. The firm found that 95% of the time, the winning vendor is already on the buyer’s day-one shortlist, and about 80% of deals are still won by the pre-contact favorite. Buyers initiate contact more than 80% of the time. By the time the conversation begins, the leaderboard is often already set.
This creates a strange and expensive ritual. Buyers are expected to research sprawling categories full of technical products they may only buy once every few years. They read analyst reports, review sites, comparison pages and vendor content written to be discoverable rather than illuminating. They try to decode overlapping claims about orchestration, identity, measurement, activation and AI. Then, after doing all that homework, they still have to run a formal process that often strips out the very context needed to make a good decision.
Which brings us to the RFI, that instrument of mutual despair.
The problem with martech RFIs is not just that they’re slow and cumbersome. It’s that they reward the appearance of certainty when there is none. A buyer turns a messy business problem into a spreadsheet of requirements. Vendors answer questions designed to reduce risk, but the exercise often obscures the real issues and can take weeks to conclude.
This being 2026, there will be some people who look at this and see another problem waiting to be solved by AI. They’ll be wide of the mark. Sure, AI can help research and summarize options. But the answer to a broken market is not to remove humans from it. It is the opposite.
Martech buying needs to get back to the fundamentals of what a market is supposed to do: connect serious buyers with qualified vendors. That means fewer templated pitches and zombie RFIs. It means creating platforms and venues where buyers and vendors can connect like adults and have genuine productive interactions that don’t waste anybody’s time.
Because the real scandal in martech is not that there are so many tools but that the system around them makes it so hard to choose wisely.
###
Sean Simon is Co-Founder of Blurbs, a vendor analysis and management platform for martech vendor discovery. He is also host of The MarTech Matrix podcast.
