Quick Answer: In Setup's 2025 survey, clients mostly said they left because of dissatisfaction with delivery and value, while agencies mostly blamed client budget cuts. Retention is helped by showing progress against the client's goals, explaining results in plain language, and always giving a next action. The survey is small, so treat the numbers as indicative.
Agencies often assume clients leave over price, or because results dipped. The data we can find points somewhere else.
This post uses primary sources where we could reach them, says how reliable each one is, and turns it into steps. We sell reporting software, so keep that in mind and check the sources yourself.
What clients say: Setup's 2025 survey
Setup, which advises brands on agency relationships, publishes an annual Marketing Relationship Survey. Its 2025 results were produced with the CMO Institute. According to Setup, there were roughly 100 true responses from agencies and brands, 57% from clients and 43% from agencies.
Reasons clients gave for ending agency relationships in 2025:
- Dissatisfaction with delivery: 61%
- Dissatisfaction with value: 61%, up from 36% in 2024
- The agency did not understand their business: 44%
- Dissatisfaction with the relationship: 41%
- Dissatisfaction with strategic approach: 41%
- Budget cuts: 39%
- Change in client leadership: 35%
- Change in agency personnel: 28%
What agencies think: the delivery gap
The same survey asked agencies why they think clients leave:
- Client budget cuts: 75%
- Change in client leadership: 55%
- Change in agency personnel: 30%
- Client dissatisfaction with delivery: 18%
- Agency did not understand the client's business: 3%
That is the gap: 61% of clients cite delivery, while 18% of agencies think delivery is the reason. Setup's multi-year roundup shows the client figure for delivery at 48% in 2024 and 61% in 2025.
How reliable is it? About 100 responses is a small sample, and the participants are people who chose to answer a survey from an organization that works on agency selection. It shows a pattern worth taking seriously, not a precise rate. Some vendor blogs quote an older version of these numbers, including a 48% figure labeled 2025, but Setup's own pages place that 48% in 2024.
What the churn benchmarks say
Focus Digital, an agency, published an Average Marketing Agency Churn report on July 6, 2026, based on research from January to April 2026.
By business model, annual churn:
- Retainer-based: 18% (average client lifespan 56 months)
- Hybrid: 28% (36 months)
- Performance-based: 33% (30 months)
- Project-based: 42% (24 months)
By service line:
- Full-service digital: 25%
- SEO services: 38%
- Paid advertising (PPC): 49%
By agency size:
- 1 to 10 employees: 32%
- 11 to 25: 24%
- 26 to 50: 19%
- 51 or more: 15%
How reliable is it? The report does not state its sample size and does not define how it calculates churn. It also says project-based churn is not always dissatisfaction, because high turnover often reflects projects naturally ending. We also noticed that some figures in its narrative sections differ from its tables. Use these numbers to see roughly where you sit, not to set a target.
Numbers you may see quoted that we could not verify
Two sets of statistics circulate on agency blogs, and we left them out of this post:
- A claim that 80% of clients who leave agencies cite poor communication, attributed to an UpCity survey. We could not find the original survey
- A set of figures (68% lack of proactive strategic guidance, 57% poor communication, 53% inability to show value) attributed variously to Havas or to GigRadar research. We found that Havas's actual client-agency research is a different study, and we could not find an original source for these numbers
If you cite either, find the original first.
What reporting has to do with it
Two claims show up repeatedly from tool vendors:
- Reporting matters a lot. Swydo's retention article calls reporting the single biggest retention activity by impact. That is the author's opinion, not a measured result
- Static reports cannot answer follow-ups. WhatConverts argues a monthly PDF cannot answer a client's next question, which shakes their confidence. WhatConverts sells client portals, so it is arguing for its own solution
Neither is proof. They are plausible and fit the survey picture: if the top complaints are delivery and value, then showing progress against the client's goals and explaining results clearly addresses both. We have not found a study that shows reporting alone changes churn.
Six things you can do
These follow the survey findings and common advice.
- Report against the client's own goals. Leads, bookings, sales, not just clicks and impressions
- Explain results in plain English. Define terms and lead with the conclusion
- Always end with a next action. A client should finish the report knowing what happens next
- Be honest about weak months. See how to tell a client it was a bad month
- Catch problems before they do. Check each account on a schedule, not only when a report is due
- Know who is actually reading. If a client has not opened the last three reports, that is a signal worth acting on
How LedgeSpace fits
LedgeSpace is built around parts of this list:
- Reports are scheduled per client and sent as a PDF plus a web link, with your branding and agency commentary blocks for your own explanation
- Client Pulse shows which clients opened or viewed their reports, and lets you send a one-click follow-up to those who did not
- RUSHA, the optional AI add-on, runs background scans that compare each client's last 7 days with their usual baseline, so you can see a problem early. Anything a client sees is a draft you approve first
It does not change why clients leave on its own. If delivery is the real problem, software will not fix that.
Start a 14-day free trial with no credit card, or see how Client Pulse works.
