All articles
Operations7 min

Auditing your tech stack: which tools can go

Small marketing teams usually pay for more software than they use. Here's how to cut your stack back to what works, without breaking anything.

Ask a marketer running a team of three to list every tool the company pays for. The first ten come easily. After that, the searching starts. A scheduling tool nobody can name the owner of. A heatmap subscription bought for one website redesign. Two separate email tools, because sales once wanted their own. The invoices keep coming.

This is rarely carelessness. It's the result of years of separate decisions that each made sense at the time. Adding a tool costs fifty euros a month and solves a problem today. Removing one costs an afternoon of digging and delivers nothing today. So the stack grows and never shrinks.

The issue isn't only money. A bloated stack costs attention. Every tool has a login, an owner, a data flow and an update that eventually breaks something. For a compact team, that's the most expensive line item.

Start with the bank statements, not the tools

Most inventories begin with a list of tools you can think of. That doesn't work, because you'll forget exactly the subscriptions you're trying to find. Start with finance instead.

Request an export of all recurring charges from the past twelve months. Filter for anything that smells like software. Do this for the company account and for the credit cards held by marketing, sales and management. Annual invoices hide for eleven months at a time.

Add two other sources: the list of connected apps in your Google Workspace or Microsoft 365 (that's where the free accounts live that still pull data), and the scripts running on your website through your tag manager. Tags nobody recognises are almost always leftovers.

The result is usually bigger than expected. Not because anyone did something wrong, but because nobody had ever put it in one place.

Four questions per tool

Put the list in a spreadsheet and answer four questions per row. No more, or it becomes a project that never finishes.

  1. 1Who has opened this tool in the past thirty days? Not 'who owns it', but who was actually in there. Most tools show this under user management.
  2. 2What decision or action comes out of it? If the answer is 'we look at it sometimes', that's an answer.
  3. 3Does it hold data we don't have elsewhere? A tool that only presents what's already in your analytics or CRM is a duplicate.
  4. 4What happens if we switch it off tomorrow? Sometimes a form breaks. Sometimes nobody notices.

Question four matters most and is the hardest. Tools are connected more often than you'd think. A forgotten automation tool can be the only bridge between your forms and your CRM. So: map first, cancel later.

Where things usually go wrong

The same patterns keep showing up. These are the places where an audit most often turns up overlap or dead weight:

  • Overlapping analytics. Your web analytics, your ad platforms and a separate dashboard product all measure conversions, each with its own definitions. The result is arguments about numbers instead of decisions based on them.
  • Project tools next to project tools. One task board for marketing, another for the agency, and loose agreements in a chat tool. Work disappears precisely between the systems.
  • One-off research tools. Heatmaps, surveys, A/B testing software. Often bought for a single project and paid for ever since.
  • Content libraries and stock. Frequently multiple subscriptions, sometimes in the name of someone who left the company.
  • AI tools from the past two years. This is where growth is fastest. Separate credits with five providers while one platform covers most of it.
  • Tools belonging to departed colleagues. The account sits on a personal email address. Nobody can access it, the charge keeps running.

Cancelling without breaking things

Cleaning up a stack goes wrong when you cancel ten subscriptions in one week. Work in rounds, with a pause in between.

First, stop everything nobody has logged into for three months and that has no script or integration attached. That's the safe round. Next: tools with overlap. Pick one, move the data, set a cancellation date thirty days out for the other. Only then the borderline cases.

Two practical points. Always export your data before cancelling; once the subscription ends it's often gone. And put the cancellation in the calendar with the right date, because many contracts renew silently and then you've lost another year.

—A tool earns its place not because it works, but because someone makes a weekly decision with it.

What's left, and why that's enough

A compact B2B marketing team generally needs a small core: somewhere contacts and deals live, somewhere content lives, something to measure with, something to send with and something to divide work with. On top of that, depending on your model, the ad platforms and one or two tools that do something the rest can't.

What that means in practice depends on your situation. A SaaS company with product data and a long funnel has different requirements than a recruitment organisation steering on volume and speed. Anyone claiming there's one correct stack hasn't looked at your situation.

One caveat. Fewer tools is not a goal in itself. If you cancel something and lose an important measurement or turn a process manual, you've moved cost into hours. Count that in.

Keeping it clean

An audit you run once gives you a snapshot. Within a year, new things will have crept in. Three agreements that limit this, and cost little time:

  • One list with every tool, its owner, cost and renewal date. A spreadsheet is fine.
  • Adding a new tool means someone puts their name on it and notes when you'll evaluate.
  • Twice a year, one hour to run through the list, preferably just before your budget round.

In engagements where we work on performance with a small marketing team, this is often the first step. Not because saving money is the point, but because it's hard to steer on numbers that come from four systems with four definitions. Block an afternoon, pull up the charges and build the list. At least you'll know what you're paying for. What you cancel after that is an easier conversation than it looks.

Where is your growth stalling?

Half an hour to go through your situation and an honest answer on whether REX.digital can help.