Summary
- Every agency switch resets months of onboarding and learning. In our estimate, three switches in four years can cost well over a year of momentum.
- Most carousels start inside the company: a vague brief, ownership split across vendors, and a scorecard nobody agreed.
- In-house leaders face the same churn: average Fortune 500 CMO tenure is 4.3 years, below the 4.9-year C-suite average (Spencer Stuart).
- Before you switch again, fix the three things you control: the brief, who owns the outcome, and what success means.
Written for: CEOs, founders and marketing heads of mid-size companies on their second or third marketing partner in a few years.
Why do mid-size companies keep switching agencies?
Because the agency is the most visible thing to change, and the easiest. The underlying problems (an unclear position, fragmented ownership and a scorecard nobody agreed) travel to the next agency intact. Founders describe this constantly. While researching this piece we read one post about paying an agency $5,000 a month for a “harsh lesson”, and another from a founder who had “tried 3 different marketers/agencies” and was still stuck at the same revenue.
The churn isn’t unique to agencies. Spencer Stuart’s 2025 study put average Fortune 500 CMO tenure at 4.3 years, up from 4.2 a year earlier but still below the 4.9-year average across the C-suite. And Gartner found only 27% of CEOs and CFOs said their CMO exceeded expectations. When expectations are vague, whoever runs marketing, inside or outside, tends to be replaced before the work compounds.
Source: Spencer Stuart
What does each switch actually cost?
More than the fees. Every new partner spends its first months learning your business, redoing what the last one did differently, and rerunning experiments you’ve already paid for. Meanwhile the market sees three versions of your company in four years.
| Hidden cost | What it looks like | How big it gets |
|---|---|---|
| Onboarding | Workshops, audits and access requests that repeat what the last agency already learned | Weeks of leadership time, every time |
| Lost learning | Test results, audience insight and keyword history that leave with the old team | Months of experiments, paid for twice |
| Rework | A new website, new templates, a “brand refresh” so the new partner can put its stamp on things | Often the largest line |
| Inconsistency | Buyers see a different voice, look or promise every eighteen months | Hard to price, easy to feel in win rates |
| Leadership trust | Each failed partner makes the next marketing proposal harder to approve | Compounds with every switch |
Rework is the cost founders underestimate most. One post we came across described a team that had rebuilt its website seven times in 56 days. That’s extreme, but the instinct is common: when results disappoint, rebuild the most visible thing.
- One partner, one plan
- A new partner every 18 months
Source: 3Anomaly illustrative model
Is it the agency, or is it the brief?
Test the brief first. Four signs that the problem travels with you:
- The brief asks for “more leads” or “brand awareness” without a number, a segment or a deadline.
- Brand, website, performance and content sit with different vendors who never meet.
- Nobody agreed, in writing, what a qualified lead is before the first report.
- The founder approves every creative decision but doesn’t attend the monthly review.
If two or more of those are true, a new agency will hit the same wall. If none are, and results still haven’t moved after two quarters against an agreed scorecard, the partner may well be the problem.
How do you get off the carousel?
Write a brief that could survive an agency change
One page: the business goal in rupees, the segments that matter, the position you want to own, the metrics and their definitions, and what you won’t do. If a new partner could start from it without a discovery workshop, it’s good enough.
Give one team ownership of the outcome
Fragmentation is the most common driver of the carousel. When brand, website and growth sit with different vendors, each hits its own KPI while the business result belongs to nobody. Whether it’s one partner or an internal lead coordinating several, one name has to own the number.
Agree the scorecard before the work starts
Qualified pipeline, cost per qualified opportunity and win rate by source, with definitions signed by finance and sales. Our piece on marketing reports that survive the boardroom sets out the five numbers we’d use.
Commit to two quarters, then decide on evidence
Set a review at six months, against the scorecard agreed on day one. Replace partners on evidence, not frustration, and make sure you own everything they built: accounts, data, files and documentation.
What should the contract say?
Enough that the next switch, if it comes, doesn’t cost you the work. Most carousel damage happens at exit: accounts held in the agency’s name, no documentation, a website nobody else can edit. Six clauses prevent most of it.
| Clause | What it says | Why it matters |
|---|---|---|
| Ownership | All accounts, data, files, code and content are the client’s, set up in the client’s name from day one | You keep the asset, whoever runs it |
| Documentation | Strategies, test results and playbooks are delivered in writing as the work happens | Learning stays with you, not in a departing team’s heads |
| Scorecard | The metrics, definitions and review dates are part of the scope of work | Nobody can redefine success after the fact |
| Named people | The senior people who pitched are named, with notice if they change | You get the team you chose |
| Transparent time | Hours or deliverables are itemised every cycle | You can see where the money goes before the invoice |
| Handover | A defined notice period with a handover plan and full access transfer | Exits become orderly, not hostile |
If the brief doesn’t change, the next agency won’t either.
Questions to take into your next leadership meeting
- How many marketing partners have we had in four years, and what did each leave behind that we still use?
- Could a new partner start from our brief tomorrow, without a discovery workshop?
- Who owns the business outcome of marketing: one name, or several vendors?
- Did we agree what a qualified lead is before the last partner started?
- What would we need to see in six months to keep the next partner, rather than replace them?
Questions leaders ask us about this
How long should we give a new agency?
For growth work, two quarters against a scorecard agreed on day one is usually fair. Brand and website projects should be judged on delivery against the brief. Any shorter and you’re judging onboarding, not performance.
Is one integrated partner better than specialists?
Not automatically. What matters is that one person or team owns the outcome and everyone works from the same brief and position. An integrated partner makes that easier; a strong internal lead can achieve it with specialists.
What should we keep when we switch?
Everything: account access, analytics, raw design files, test results, keyword and audience data, and documentation. Make your ownership of these a contract term from day one, not a negotiation at exit.
Expert verdict
Fix the brief, the owner and the scorecard before you fire anyone.
Agencies do fail, and some deserve to be replaced. But when a company has churned through three, the pattern is usually internal: no written position, ownership split across vendors, and success defined after the fact. Fix those and even an average partner improves. Skip them and the best agency in Mumbai will be gone in eighteen months too.
Do now
Write the one-page brief and the scorecard, and share both with your current partner before deciding anything.
Stop
Splitting brand, website and growth across vendors with no single owner of the result.
Measure
Qualified pipeline and cost per qualified opportunity, reviewed at a fixed six-month point.
Verdict by our Designing lead, 3Anomaly founding team
