How to Handle New Martech Requests Without Rebuilding the Sprawl

By Antonio Caruso, Caruso Martech

Published Nov 5, 2026 · Updated Nov 5, 2026 · Automation & Intelligence

Every quarter someone wants a new tool added to the stack. A short request process keeps the good ones coming in and stops the sprawl you just cut from growing back.

Every quarter someone in marketing wants another tool added to the stack. A webinar platform, an AI writing assistant, a new reporting layer. Each request sounds reasonable on its own, and the person who looks after the stack ends up saying yes to most of them because there is no agreed way to say anything else.

That is how a stack you cleaned up last year quietly doubles again. Zylo's 2026 SaaS Management Index puts the average company portfolio at 305 applications, and marketing is one of the teams that keeps adding to it. A request process gives you a fair, fast way to decide which new tools earn a place.

Quick answer: how do you handle a steady stream of new martech requests?

  • Route every request through one short form that starts with the problem, before any vendor name.
  • Check whether a tool you already pay for can do the job.
  • Give every approved tool a named owner and a success measure before purchase.
  • Time-box new tools with a 60 or 90 day trial and a review date.
  • Review the whole stack once a quarter and retire what nobody uses.

Why requests keep coming in

New tool requests keep arriving because buying software has become easy and nobody owns the decision. Most tools start on a card with a free trial. By the time anyone in operations hears about them, the team relies on them and removing them feels like a fight.

The data shows how far purchasing has moved away from central teams. According to CFO Dive, Zylo's 2023 index found IT led only about 18 percent of software purchases, while employees bought 37 percent of apps and business unit leaders 45 percent.

That spread is a governance gap. The people buying tools are trying to solve real problems, and the fix is giving them a route where those problems get solved without every one becoming a new subscription.

The cost of saying yes by default

Saying yes by default costs more than the subscription. Each new tool adds a login, an integration to maintain, data that has to match the CRM, and another thing to train new hires on. Those costs land on whoever runs marketing operations, usually without any budget attached.

Unused capacity is the clearest sign. Zylo's 2024 index found companies were only using 49 percent of their provisioned licences, and that more than a third of their applications were shadow IT.

We see the same pattern in small teams on a smaller scale. Two tools doing the same job, a third one nobody remembers buying, and an operations lead who spends Fridays fixing the sync between them. That load is why we argue marketing ops is so often underfunded.

The request process, in five steps

A workable request process has five steps, and it should take days. It starts with the problem, checks the existing stack, names an owner and a measure, runs a time-boxed trial, and ends with a review date. Each step removes a common reason tools get bought and then abandoned.

  1. The requester fills in a one-page form that describes the problem, the people affected and what happens today, with any vendor name added last.
  2. The stack owner checks whether a tool you already pay for can solve the problem, using your current inventory.
  3. If a new tool is needed, the request names an owner who will run it and a measure that shows whether it worked.
  4. The tool starts on a 60 or 90 day trial with a fixed review date, and security and finance see it before any annual contract.
  5. At the review date, the owner shows the measure, and the tool is either kept, replaced or cancelled.

Step 2 is where most requests resolve. In our experience, a large share of "we need a new tool" requests turn out to be features the team already has and never set up. Others need a small automation between two tools you already pay for.

Who decides, and how fast

One person should own the decision, and the decision should come back within a week. A process that takes a month teaches people to go around it. In a small team the owner is usually the head of marketing or the marketing ops lead, with finance consulted on anything above an agreed spend.

Ownership matters after purchase too. A MarTech conference panel on martech buying recommended identifying an internal owner of every tool and bringing information security in much earlier in the process.

Speed is what keeps people inside the process. If a request gets a clear answer in five working days, most people will use the form, and the owner gets an early view of what the team is struggling with. If it disappears into a queue, the next tool goes on a personal card instead.

What to do with the requests you turn down

A declined request still points to a real problem, and it deserves an answer. Tell the requester what the team will do instead, whether that is configuring an existing feature, adding it to the next quarterly review, or accepting the workaround for now. A plain "no" just sends the purchase underground.

The CFO Dive piece makes the same argument for software in general: banning employee purchases outright rarely works, and companies do better with controls and an internal policy on how apps get bought. Zylo calls this freedom within a framework.

Keep a short log of declined requests and the reason for each. When the same problem comes up three times, that is a signal to look for a proper fix at the next stack review.

The quarterly review that stops sprawl growing back

A quarterly review closes the loop. Once every three months, list every tool, its owner, its cost, and whether it hit the measure set at approval. Anything without an owner or a result goes on a shortlist to cancel at renewal. It takes an afternoon and keeps the stack honest.

If you have never done a full pass, start with a one-day stack audit to build the inventory the request process depends on. After that, the quarterly review only has to check what changed.

The review is also where the request process proves its value. Teams that have just been through a consolidation usually find the request form is what keeps the savings from disappearing within a year.

If new tool requests keep landing on one overloaded person, we help teams set up the request form, the inventory and the quarterly review in a few weeks. See our services or contact us to talk it through.

Caruso Martech

We write about marketing systems, attribution, and growth operations because these are the problems we work on every day. If something in this post is relevant to what you're building, we're happy to talk through it.

Related insights