Why We Left ClickUp: Cutting Agency PM Bloat

We ran ClickUp for 14 months across 22 client retainers. It didn't fail. That's the uncomfortable part — nothing broke, no data was lost, no client noticed. We left because the tool was quietly consuming about six hours of senior time a week and returning less clarity than a spreadsheet would have.

This is a specific account of what PM bloat cost us, how we measured it, and the process we used to cut it. If you're an agency owner staring at a workspace nobody fully understands anymore, this is for you.

What PM bloat actually is

PM bloat isn't "too many features." Every tool has features you don't use, and that's fine — nobody complains that Excel has array formulas. Bloat is when configurability becomes an unpaid job.

The mechanic works like this. A flexible PM tool asks you to make hundreds of small decisions: which statuses per list, which custom fields per space, which automations fire on which triggers, which of eleven view types each client folder defaults to. Every decision is reasonable in isolation. Collectively they create a system that only its architect understands — and agencies rarely have a dedicated ops architect until they're past 25 people.

So the system drifts. One PM builds a beautiful client space with 14 custom fields. She leaves. Her replacement doesn't know why "Content Stage" and "Draft Status" both exist, so he uses neither and tracks progress in a Slack thread. Now you have two sources of truth and one of them is unsearchable.

Our actual numbers

Before we made any decision, we spent three weeks logging where PM time went. Not project time — tool time. The results:

• 4.5 hours/week of admin re-tidying: fixing statuses, merging duplicate tasks, reassigning things filed in the wrong space.
• 2 hours/week of my time answering "where should this live?" in Slack.
• 6.5 hours average onboarding per new hire, spread over their first two weeks, before they could confidently create a task in the right place.
• 31% of tasks created in the previous 90 days had no assignee, no due date, or both.
• Three parallel systems in active use for the same information: ClickUp, a "Client Status" Google Sheet the account team actually trusted, and Slack channels per client.

That last one is the tell. When a team builds a spreadsheet alongside a £2,400/year PM tool, the spreadsheet is telling you what the tool failed to do. Ours had eight columns: client, phase, hours used vs. retainer, next deliverable, blocker, owner, last client contact, health. That's the whole job. The PM tool had 200 fields and couldn't produce that view without a paid dashboard add-on and forty minutes of setup.

The three symptoms that made the decision for us

1. Retainer reality lived outside the tool

We sell monthly retainers between £2,500 and £9,000. The single most important number in the business is hours consumed against hours sold, per client, mid-month. ClickUp tracked time. It did not natively tell us "Client X is at 78% of their retainer on the 14th" without stitching together a rollup field, a time report, and manual entry of the monthly allowance.

We found out we'd overserviced one client by 41 hours over a quarter — roughly £3,700 of unbilled work — because the signal wasn't visible where the work happened. If you're not sure whether this is happening to you, the warning signs are well documented; we later recognised four of the seven in our own book of business (how to spot unprofitable retainers before they drain you covers the pattern properly).

2. Onboarding cost exceeded the licence cost

At 6.5 hours of onboarding per hire and roughly 30% annual churn on a 16-person team, we were burning around 32 hours a year just teaching people our workspace conventions. At a blended £55/hour that's £1,760 — more than the software. And that assumes the conventions were documented, which they were, in a 4,000-word Notion doc nobody read past the second heading.

3. Client-facing views required a rebuild every time

Clients don't want a Gantt chart of subtasks. They want: what did you do, what's next, are we on track, what do you need from me. Producing that meant either giving them guest access to a view we'd sanitise manually, or exporting to a deck. We did the deck. Every month. For 22 clients. That's the definition of a tool not doing its job.

How to run the decision properly

Don't start with tool comparisons. Start with a requirements list you derive from your own behaviour, because your workarounds are the requirements.

Step 1: Audit before you shop

Spend two weeks logging tool time and cataloguing every spreadsheet, Slack channel, and Loom that exists because the PM tool couldn't do something. We ran this as a full stack audit and found £680/month of overlapping subscriptions in the process — the method is worth doing annually regardless of whether you switch (here's how to run a SaaS stack audit).

Step 2: Write the ten-item requirement list

Cap it at ten. Ours, in priority order:

1. Retainer hours consumed vs. sold, visible per client without configuration
2. Time tracking at task level with under-5-second entry
3. Phase-based project structure (audit → fix → content → link → report)
4. Client-readable status view with no manual sanitising
5. Recurring monthly retainer templates that instantiate automatically
6. Capacity view per person for the next 3 weeks
7. Under 60 minutes to full competence for a new hire
8. Task-level profitability rollup
9. Native technical SEO audit tracking (not a generic checklist)
10. CSV export of everything, in case we're wrong again

Notice what's absent: dependencies, portfolio dashboards, custom automations, goal OKR trees, form intake, whiteboards, mind maps, docs. We'd used exactly two of those in 14 months. Cutting them from the requirement list is what made the shortlist short.

Step 3: Test with one real client, not a sandbox

Sandboxes flatter every tool. We ran a single mid-size retainer in parallel for four weeks — real deliverables, real time entries, real monthly report. Two of our three shortlisted tools fell out in week two because building the retainer template took longer than the work it managed. If you're at this stage, the comparison hub is a faster way to eliminate options than booking five demos, and it's worth reading the specific PeakKR vs ClickUp breakdown if ClickUp is your incumbent.

Step 4: Migrate active work only

We exported 14 months of history to CSV, dropped it in cold storage, and never opened it. Migrating only the ~340 open tasks across 22 clients took nine working days. Rebuilding archived history would have taken a month and served nobody.

What actually got better — and what didn't

Honest accounting, six months on:

• Admin re-tidying dropped from 4.5 to about 1.2 hours/week. That's roughly 170 hours a year recovered.
• New-hire time-to-competence went from 6.5 hours to under an hour, mostly because there were fewer choices to explain.
• The Google Sheet died in week three. That mattered more than any metric.
• We caught two retainers trending toward overservice in month one and adjusted scope before the invoice — the kind of mid-month signal that makes retainer renegotiation conversations factual instead of awkward.

What didn't improve: we lost some genuinely useful automation. ClickUp's automation builder was better than what we moved to, and we now do two things manually that used to fire on a trigger. That costs maybe 20 minutes a week. We accepted the trade because 20 minutes of known manual work beats four hours of unknown drift.

And a warning: a simpler tool doesn't fix a process problem. If your team doesn't log time, no tool makes them. We had to do separate work on that — the resistance is almost always about being judged, not about the interface (tracking time without micromanaging is the piece that helped us most). Two of our four PM problems were cultural. Switching tools solved the other two and made the cultural ones visible, which is a decent outcome but not the one the sales page promises.

Should you leave ClickUp?

Not necessarily. Stay if you have a dedicated ops person who owns the workspace, if you genuinely use dependencies and automation daily, or if your projects are complex enough that flexibility beats speed. ClickUp is a good tool badly