Why Good Agencies Fail and Bad Agencies Survive
Most agency relationships do not collapse because of a single catastrophic mistake. They erode slowly through three failure modes that owners rarely name until the money is gone. The first is activity reporting instead of outcome reporting: you get a monthly deck full of impressions, clicks, and posts published, while the number that actually matters — booked jobs at an acceptable cost — sits somewhere nobody is looking. The second is channel-silo blame: ads blames the website, SEO blames the intake team, the website blames the leads, and the owner pays for a circular firing squad. The third is vanity metrics standing in for performance: traffic is up, followers are up, and revenue is flat. The scorecard below exists to make these failure modes visible early, while you can still fix them.
It works in three contexts: evaluating an agency before you sign, auditing one you already have, or scoring your own in-house marketer. The framework is identical because the mechanism is identical — someone is spending your money to produce customers, and you need a shared definition of what counts.
The Scorecard: Six Dimensions That Actually Predict Performance
1. Lead quality and booked jobs
The only honest metric is the one that survives contact with your dispatch board. Every lead source should be scored by what happened after the call: booked, quoted, won, lost, junk. An agency that cannot tell you their share of booked jobs is reporting on their effort, not your result.
2. Cost per acquired customer trend
Not cost per lead — cost per customer, trended over time. Early months will be volatile; the question is whether the trend bends toward efficiency as targeting, creative, and landing pages mature. A flat or worsening trend with rising spend is a warning regardless of what the click-through rate says.
3. Pipeline visibility
Can you open your CRM and see every marketing-sourced opportunity, its stage, and its value right now? If the agency's reporting depends on their spreadsheet and your data lives in their account, you have a visibility problem that will become an exit problem.
4. Speed of iteration
Strong operators ship weekly: new ad copy, new landing page tests, refined call handling. Ask how many substantive experiments ran last month. An agency that has not changed anything meaningful in 60 days has stopped working and started billing.
5. Communication cadence
A weekly written update of a few lines beats a monthly hour-long call. You want a predictable rhythm: what shipped, what the numbers say, what happens next. Silence between invoices is the loudest red flag in the industry.
6. Who owns tracking truth
There must be one named person — on your side or theirs — responsible for the accuracy of call tracking, form attribution, and CRM source fields. When tracking breaks (and it will), someone must catch it within days, not discover it in a quarterly review.
Questions to Ask Before Signing
The contract negotiation is where accountability is won or lost. Ask, in writing:
- Who owns the ad accounts, GBP, analytics, and call tracking numbers? The answer must be you, with the agency granted access. Agencies that build assets in their own accounts are building leverage over you.
- Who owns the data and the historical performance record? Include an explicit clause: full data export on request, at any time, in usable formats.
- What are the exit terms? Notice period, handover obligations, whether anything pauses or gets deleted if you leave. Thirty days' notice with complete access transfer is a fair standard.
- What does the first 90 days look like, concretely? A serious agency can describe the onboarding sequence without hesitating.
- How do you report on booked revenue, not just leads? If the answer is "we don't have visibility into that," ask how they expect you to judge them.
Red Flags vs. Green Flags
Red flags: reports lead with impressions and clicks; access owned by the agency; blame directed at your other vendors instead of proposed tests; guaranteed rankings or lead volumes; long contracts with no performance review clause; you cannot name a single experiment from last month.
Green flags: they ask about your close rate and average ticket before quoting; they volunteer account access terms; bad months come with a diagnosis and a plan, not excuses; their first deliverable is a measurement plan, not a media buy; they talk about your CRM as their workspace.
The First 90 Days: What Onboarding Should Look Like
Days 1–30 are for truth. Baseline every number: lead volume by source, booked-job rate, cost per customer, close rate, average ticket. Fix tracking before scaling spend — call tracking, form source fields, CRM stages. Nothing should be "optimized" against broken measurement.
Days 31–60 are for first signal. Campaigns launch or get restructured, landing pages get tested against the current ones, and the first full-funnel report exists: spend to lead to booked job to revenue. Expect noise, not miracles — the goal is a trustworthy feedback loop.
Days 61–90 are for the trend. You should see early efficiency gains, a concrete iteration backlog, and a clear statement of what the next quarter's bets are. By day 90, hold a formal scorecard review against the six dimensions. Decide to scale, adjust, or exit — but decide on evidence, not vibes.
Running a Monthly Review That Produces Decisions
A productive review is short, numerical, and ends in commitments. Thirty to forty-five minutes, same agenda every month: booked jobs and revenue by source; cost per acquired customer trend; pipeline movement; what shipped and what was learned; the two or three decisions this month. The review is not a status update — it is a decision meeting. If a metric moved, name the mechanism and the next action. If nothing can be decided, the reporting is too thin.
Do This Now
- Pull last month's agency report and circle every number that connects spend to booked revenue. If you can only circle leads and clicks, you are buying activity.
- Log into your ad accounts, analytics, GBP, and call tracking today. If you lack admin access, request it this week.
- Score your current provider (or in-house marketer) against the six dimensions on a one-to-five scale. Anything at three or below gets a named fix and a deadline.
- Draft the three exit-term questions and send them before your next renewal conversation.
- Put a recurring monthly review on the calendar with the agenda above, and make attendance conditional on the full-funnel report arriving 48 hours early.
Accountability is not adversarial — the best agencies welcome it, because a client who measures outcomes clearly is a client worth doing great work for. The scorecard simply makes sure both sides are playing the same game: customers acquired at a cost your margins can carry, with nothing hidden in the reporting.