Nine Subscriptions. Four Tools You Actually Use.

The Renewal You Skim

Once a year somebody in your agency opens the card statement, sorts it by vendor, and discovers a $49 line item nobody has logged into since March. It gets canceled. Everyone feels efficient. The stack is exactly as expensive as it was the day before.

That is the shape of most tool audits: they find the dead subscriptions and miss the expensive ones. The $49 tool nobody uses costs you $588 a year. The $200 tool two people use badly, that nothing else talks to, and that produces the report a client argues with every month — that one costs multiples more, and it never shows up as a line worth canceling because somebody is clearly using it.

Agencies are unusually exposed here. According to Basis’s 2026 Advertising Agency Report, 36.8% of full-service and media agencies now manage ten or more tools — more than double the 17.3% recorded two years earlier. The pattern is not agency-specific either: Zylo’s 2026 SaaS Management Index puts the average organization at 305 applications, up from 275 the year before, including 9.9 project management apps and 9.5 team collaboration apps running side by side.

Nobody chose that. Stacks are not designed, they are accumulated — one tool per crisis, one per client demand, one per new hire who brought a favorite with them. This is how you take the accumulation apart.

Three Costs, One Invoice

A tool bills you once and charges you three times. Only the first charge appears on a statement.

1. Subscription

The visible one. Easy to count, easy to negotiate, and almost always the smallest of the three. Note that seat-priced tools have a second-order cost: they tax you for growing. Any tool where adding a junior designer adds a line to the invoice is a tool that quietly argues against hiring.

2. Swivel-chair labor

The work a human does purely because two systems do not talk. Exporting a CSV from the ad platform to paste into the reporting deck. Re-typing a lead from the form tool into the CRM. Copying an approved caption from the doc into the scheduler. None of it is billable, none of it is skilled, and all of it is invisible until you count it.

Price it honestly: minutes per week × 52 × the loaded hourly cost of whoever actually does it. Not your rate — the rate of the account coordinator who really does it at 5:40 on a Friday.

3. Reporting drift

The most expensive and the least measured. When four tools each hold a version of the truth, your monthly client report is a reconciliation, not a readout. Sessions in one place, conversions in another, spend in a third, and a number in the deck that matches none of them exactly. The cost is not the hour spent reconciling. The cost is the client call where a number is questioned and you cannot immediately defend it.

Run the Audit

One row per tool. One hour, one spreadsheet, the whole team in the room because nobody knows the whole stack alone. The columns matter more than the precision — you are looking for shape, not accounting.

Two columns do the real work. Only it does… is where overlap becomes undeniable — when three rows all say “schedules social posts,” the argument is over. Feeds / fed by is where the hidden labor surfaces; every “manually” is a person, a recurring calendar block, and a place where numbers drift apart.

Keep, Cut, Consolidate

With the sheet filled in, every row resolves into one of three buckets. Work them in this order, because cutting is free and consolidating is not.

Cut

• No named owner, or the owner has left.
• The “only it does…” cell duplicates another row and the other row is better.
• Bought for a client who is no longer a client.
• Last meaningful login is older than the last invoice.

Keep

• It does one job nothing else does, and does it well.
• It is genuinely best-in-class and your team would notice a downgrade within a week.
• A client contractually requires it, or their data lives in it.
• It is the system of record for something you are legally obliged to retain.

Consolidate

• Two or more rows overlap on the same job and neither is clearly better.
• The “feeds / fed by” cell says “manually” and the volume is weekly or higher.
• The tool is fine but the seam around it is not — the export, the re-key, the reconcile.
• It bills per seat for a function the whole agency needs.

What Consolidation Breaks

Consolidation is sold as pure upside. It is not, and going in clear-eyed is the difference between a migration that sticks and one that quietly reverts within two months.

You lose history at the boundary

Most platforms import current records well and historical activity badly. Three years of notes attached to a lead, the thread where a client approved something contentious, the audit trail from a campaign that went sideways — assume it does not move. Export it, store it somewhere you control, and do that before you cancel anything.

You lose the one feature somebody loved

Every specialist tool has a depth the generalist does not match. That is a real trade, not a marketing objection. The question is whether that depth is load-bearing for client work or a preference. Ask the person who will lose it, in writing, what they use it for — you will get a straight answer and occasionally a genuine blocker.

You pay twice during the overlap

Budget for sixty to ninety days of running both. A migration that also has to hit a cancellation date is a migration that will cut corners on exactly the data you needed.

A Worked Example

A twelve-person web design studio runs eleven tools. The audit finds four rows whose only it does… cells overlap and four whose feeds / fed by cells say “manually.”

Two cuts, three keeps, six consolidations. The interesting number is not the subscription saving — it is that six rows shared one failure mode. Every one of them existed because something upstream would not hand its data to something downstream. The studio was not paying for six tools. It was paying for six seams.

What to Do This Week

None of this requires a decision yet. It requires the sheet.

When you are ready to act on the consolidate column, the shortlist writes itself: the replacement has to cover every overlapping row, close the manual seams between them, and not reintroduce per-seat billing on the way in. That is the brief the nine modules behind one login are built to answer, and what seven client-facing surfaces on one workspace looks like in practice. If the seams you counted are mostly integrations with tools you intend to keep, start at the live integration directory instead — consolidation and connection solve the same problem from different ends.

Then run the other half of the audit. This one prices the money going out; the hours audit prices the capacity going out, and agencies at capacity usually find the second number is the one holding them back.