When an improvement plan is the right move
An improvement plan suits a relationship where the agency has fallen short on things it was asked to do, the shortfall has been raised in reviews without lasting change, and the business would rather fix the relationship than replace it. Replacing an agency costs time, knowledge and momentum; the page on results that dropped after changing agency shows how much can be lost in a handover. A plan that works keeps what the agency already knows.
It is the wrong move in three situations. If the scope no longer matches what the business needs, the fix is a rescope at the next review, which the tutorial on reviewing an agency before renewal covers. If the business has already decided to leave, a plan only delays an honest conversation. And if the problem is that the agency reports success while sales stay flat, check measurement first: the diagnosis of agency reports that look good while sales are flat separates measurement gaps from performance gaps, and a plan built on the wrong measure fixes nothing.
The plan starts from a principle the site's agency pages return to. The agencies and partners hub notes the Institute of Practitioners in Advertising's long-standing argument that the client side shapes the quality of agency work as much as the agency does. An improvement plan that only lists what the agency must change misses half of the cause.
Before you start
Gather the record before any conversation about a plan.
- The scope and the agreed measures. What the agency was asked to deliver, how success was defined and the source of each number. If the measures were never agreed, that is the first finding.
- The review record. Scorecards from both sides, monthly reviews and the issues raised, with dates. The two-way scorecard in the guide to selecting, briefing and reviewing agencies is the most useful single document.
- Results from the business's own data. Each KPIA metric tied directly to an objective. in the scope, read from the business's systems for the period in question and the period before.
- The briefs. The campaign briefs and each Creative BriefA short document telling the people making creative work what it must achieve, for whom, with which message and within which limits. the agency worked from, with the feedback and approval history.
- Channel-specific checks. For paid search, the guide to hiring and evaluating a Google Ads agency lists the account ownership, conversion definitions and reports to check; for search engine optimisation, the guide to hiring and evaluating an SEO agency lists the warning signs Google publishes. Both apply to an agency already in place.
- Your own agreement. Read it for anything it says about performance, reviews or ending the relationship, and take any question about it to the business's legal advisers before the plan is written.
The steps
Write down the gap against the scope and the agreed measures
State the shortfall in writing, specifically. For each problem, record what the scope asked for, what happened, which measure shows it and where the evidence sits: missed deliverables, late work, quality below what was agreed, or results below the baseline on the business's own data. Separate what the agency was responsible for from outcomes that depended on others. Keep opinions out of this list and examples in. A gap that cannot be written down against the scope or a measure may be a problem with the scope or the measure rather than with the agency, and the plan should say so.
Check your own share of the cause
Before talking to the agency, review the business's side. Use the client half of the two-way scorecard: were briefs clear and stable, was feedback consolidated and consistent, did approvals come in time, and did the agency have access to the results it was judged on? Read the briefs again against the campaign brief guide and the creative brief template, looking for what they left out. An agency that cannot see which leads became customers cannot optimise for customers. Write the business's share down with the same specificity as the agency's. Changed briefs and scattered feedback often show on the invoices too, as rework, and the diagnosis for agency fees that keep rising separates that share of the cost from the agency's own.
Hear the agency's account
Share the written gap with the agency's lead and ask for its account before proposing any plan. Ask what it sees as the causes, what it has tried, what it needs from the business and what it would change in the scope or the ways of working. Listen for causes the business did not see, such as late inputs from another partner or access the agency never received. Where several agencies share the work, the guide to coordinating several marketing agencies covers how problems at the joins between them are traced and settled. The aim is a shared view of the problem, not agreement on blame.
Trace the main shortfalls to their causes
Take the few shortfalls that matter most and trace each one to a cause either side can change. The five whys suits this: ask why the shortfall happened, then why that happened, until the answer is something a person can act on. "Enquiries are low quality" might trace to targeting, to the offer in the ads, to a brief that never said which services had capacity, or to AttributionAssigning conversion credit to touchpoints. that credits the wrong source. Do this with the agency in the room, so both sides see the chain and accept the cause. Shortfalls that trace to the same cause need one change, not several.
Agree changes on both sides in writing
Write the plan as a list of changes, each with an owner on one side or the other. Agency changes might cover targeting, creative rounds, reporting, staffing or process. Business changes might cover the brief, a single approver for feedback, access to results or earlier notice of launches. Keep the list short enough to act on, and make every item specific enough that both sides will agree whether it happened. Both sides sign off the plan, and both keep a copy. If any change would alter the scope or the fee, treat it as a rescope and handle the commercial side through the normal route rather than inside the plan.
Set measures, baselines and review points
For each change, name the measure that should move, where it stands today as a baseline and the source of the number, read from the business's own data. Then set the review points: the dates when both sides will meet to look at the measures and the changes, and the date the plan ends. Choose the timing from when each measure can reasonably move, which depends on the work and the sales cycle, and say in the plan what will happen at the end: the decision will be to continue, rescope or exit, taken on the evidence the reviews have gathered.
Change the account team only when the evidence points there
A new account team is sometimes the right answer and often a reflex. Ask for one only when the evidence from the earlier steps points to the people: work that falls short on a clear brief, problems that persist after the business's own changes, or specific gaps in skill or attention. Raise it with examples and let the agency propose the change. If the causes sit in the brief, the scope or the measures, a new team will inherit them, and the change will cost the knowledge the current team holds. Where the team does change, keep the measures and the review points the same so the plan stays comparable.
Hold the reviews and record what changed
Hold every review point as planned, with both sides present. At each one, go through the changes on both sides, whether each happened, and what the measures show against the baseline. Read small counts carefully and judge the trend rather than a single reading. Record the outcome of each review in a short note both sides receive: what changed, what did not, and what each side will do before the next review. Do not move the measures or the end date mid-plan unless both sides agree in writing; a plan that shifts as it goes cannot support a decision.
Decide to continue, rescope or exit
At the final review, decide on the evidence. Continue when the measures moved and the changes on both sides stuck; fold the plan's measures into the normal review rhythm. Rescope when the agency improved on what it can do but the work the business needs has changed, and use the renewal review to reset the scope. Exit when the agreed changes did not happen or did not move the measures despite the business keeping its side. Any notice or contractual step follows your agreement and the business's legal advisers. If the decision is to exit, the agency offboarding checklist covers the handover, and the agency request for proposal template starts the search for a replacement.
What the improvement plan contains
| Part of the plan | What it records | Why it matters |
|---|---|---|
| The gap | Each shortfall against the scope and the agreed measures, with evidence | Keeps the plan specific and fair |
| The business's share | Briefing, feedback, approval and access issues on the client side | Fixes the half of the cause the agency cannot fix |
| The agency's account | Its view of the causes and what it needs | Surfaces causes the business did not see |
| Root causes | The chain from each main shortfall to an actionable cause | Points the changes at causes, not symptoms |
| Changes on both sides | Each change with an owner | Makes both sides accountable for the result |
| Measures and baselines | The measure for each change, its baseline and its source | Shows whether the changes worked |
| Review points and end date | When both sides meet, and when the plan ends | Gives the plan a rhythm and a finish |
| The decision rule | What continue, rescope and exit would each look like | Removes surprise from the final review |
A dental group and its paid social agency (illustrative)
Zelquarin Dental turns round its paid social agency
Zelquarin Dental, an invented group of 6 dental practices, paid an agency to run paid social campaigns for new patient enquiries. Over 2 quarters the agency's reports showed enquiries up, but practice records showed only 9% of social enquiries became booked appointments, against 22% for enquiries from the website.
The marketing manager wrote down the gap against the scope, which asked for "new patient enquiries" without mentioning bookings. Checking Zelquarin's own side, she found the brief never said which treatments had free appointments, so the agency had promoted whitening and implant consultations at practices with waiting lists. The agency's account added that it had no view of which enquiries booked. Tracing the low booking rate with five whys led to the same cause from both directions: nobody had connected the campaigns to appointment capacity.
The plan, signed by both sides, set 3 agency changes (targeting by practice, 2 creative rounds a month with one approver, and a weekly readout of booked appointments from social) and 2 Zelquarin changes (a monthly list of treatments with free capacity, and access to the booking report). The measure was booked appointments from social, with a baseline of 31 a month, and reviews were set at weeks 4, 8 and 12.
At the final review booked appointments from social had reached 58 a month and the booking rate 19%. Zelquarin decided to continue, rewrote the scope around booked appointments, and kept the weekly readout.
Verify it worked
- The written gap lists each shortfall against the scope and a measure, with evidence and the source of the number.
- The plan lists changes for both sides, each with an owner, signed off by both.
- Every measure has a baseline from the business's own data. In the illustrative example above, the baseline was a monthly count of booked appointments from social.
- Each review has a note both sides received, saying what changed and what happens next.
- The final decision follows the evidence. In the illustrative example above, booked appointments and the booking rate had both risen by the final review, so the group continued.
If it goes wrong
- The agency disputes the gap. The shortfall is probably vague or not tied to the scope. Rewrite it against the scope and a measure.
- The two sides name different causes. Run the five whys again with the agency in the room.
- The agency's changes happen but the measures stay flat. The traced cause may be wrong, or the business's own changes did not happen. Check both lists first.
- The agency stops engaging. Hold each review as planned and record the absence. That is itself evidence for the final decision.
- Pressure builds to swap the account team. Check whether the evidence points to the people. If it points to the brief, a new team inherits the same problems.
Next steps
- If the decision is to continue, read the guide to selecting, briefing and reviewing agencies to carry the plan's measures into the two-way scorecard.
- If the decision is to exit, read the in-house versus agency comparison to weigh who should do the work next.
- Read the diagnosis of agency reports that look good while sales are flat if reported success and your own data keep disagreeing.
Common mistakes
- Skipping the client's share. Some of the cause usually sits with the brief, feedback, approvals or access. Check it before the plan, not after.
- A plan that only lists agency changes. The business's changes belong in the same document, with owners.
- Vague gaps. "Not proactive enough" cannot be measured. Write each shortfall against the scope and a measure.
- Treating symptoms. Fix the cause the five whys finds, not the first explanation offered.
- Measuring on the agency's report. Read every measure from the business's own data, against a stated baseline.
- Changing the account team by reflex. A new team inherits the same brief and the same measures. Ask for one only when the evidence points to the people.
- Moving the goalposts mid-plan. Changing measures or dates without agreement makes the final decision indefensible.
- Making contractual moves inside the plan. Notice, termination and fees follow the agreement and the business's legal advisers.
