Three months is too early to judge the outcomes of a marketing engagement on a nine-month sales cycle, and too late to judge its inputs.
That is the whole answer, and almost nobody acts on it. A founder signs an engagement in January, sees no closed revenue by April, and either fires the agency or extends it on faith. Both decisions use the wrong evidence. Closed revenue cannot exist yet when the sale takes nine months. Tracking, schedule and lead quality, the month-one checks of any growth marketing engagement, could have been checked by the end of February.
An agency operator described a client who wanted a one-month test campaign; when asked how long the typical sales cycle was, the answer was months. A one-month test on a nine-month cycle measures one thing: whether the agency can set up a campaign. It says nothing about whether the campaign sells.
So how long before you can judge a B2B marketing engagement? You judge different things at different months. Here is what a sound engagement reports at month one, three and six, the kill criterion for each, and what to put in the contract so you never have to argue about it.
Month 1 of the engagement: tracking, schedule, quality
Month one is an audit of inputs. Nothing about outcomes is knowable, and anyone presenting outcomes in month one is presenting noise.
Three things should be on the first report.
Tracking. Every enquiry, call and form is recorded, attributed to a source, and lands in the CRM with the fields sales needs. If the engagement includes paid media, offline conversion import from the CRM is set up or scheduled, so that spend will eventually be judged on pipeline stage rather than form fills. Our B2B ROAS measurement article describes what that architecture looks like when it is finished; in month one you only need to know it is being built.
Schedule. What was promised for the first ninety days, and whether the first thirty are on it. Campaigns live, pages published, the account list agreed with sales, the first content out.
Quality. Of the enquiries that have arrived, how many match the ideal customer profile. Ten leads with two in the target market is a quality problem you can see in week three. Waiting until month six to notice is the most common waste in these engagements.
Month-one kill criterion: tracking not working and no dated plan to fix it by day 45. Everything else in the engagement depends on it, and an agency that cannot get tracking right in six weeks will not get attribution right in six months.
Month one proves the engagement can measure; nothing else is provable yet.
Month 3: growth, scaling spend, one to four winners
By month three, the leading indicators move. Not revenue. Movement.
A sound month-three report shows growth in the things that precede pipeline: qualified enquiries up on month one, cost per qualified enquiry falling, a rising share of enquiries from target accounts. It shows spend being scaled into the channels, audiences or pages that are producing qualified enquiries and pulled from the ones that are not. And it names the winners: one to four specific things (a search theme, an audience, a landing page, a content angle) that are working well enough to put more behind.
One to four is the right number. Zero winners by month three means the targeting, the offer or the message is wrong, and the report should say which the agency thinks it is and what changes in month four. Ten winners means nothing is being measured tightly enough to tell.
What you still cannot see at month three on a nine-month cycle is a closed deal from the engagement's own work. What you can see is whether the pipeline stage definitions are being hit: first meetings booked from new sources, opportunities created, proposals out. Our guide to B2B sales cycle metrics covers which of those are worth tracking and which are vanity.
Month-three kill criterion: no winner identified and no hypothesis for why. Not "no revenue". A missing hypothesis is the signal, because it means the engagement is running on activity rather than learning.
Month three proves the engagement can learn; revenue is still ahead of it.
Month 6: the first losers, new winners, scaling continues
Month six is where a sound engagement starts to look like a system rather than a set of campaigns.
The report now includes losers: things that were tried, measured and stopped. An engagement with no losers at month six has either been very lucky or has not been testing. New winners appear alongside the original ones, because the first set has been scaled as far as it goes and the next layer has been found. Spend keeps moving towards what works.
Pipeline, by stage, is now the headline. On a nine-month cycle, the earliest enquiries from month one are two-thirds of the way through the sale. Some have become proposals. Some have died, and the report says why. One or two may have closed if the firm's cycle has a short tail. The report shows the pipeline the engagement has created against the spend to date, by stage, with the stage definitions sales agreed in month one.
This is also the month to check the arithmetic the agency reports against your own systems. The agency ROAS reporting audit sets out how to reproduce a claimed figure from source records in an afternoon; month six is the first point at which there is enough data to make that worth doing.
Month-six kill criterion: qualified pipeline created by the engagement is below the spend to date, and the trend from month three to six is flat. One of those alone is a conversation. Both together is a decision.
Month six proves the engagement can compound; closed revenue is still arriving.
A worked engagement timeline on a nine-month cycle
Take an illustrative services firm with a nine-month average sales cycle, a £40,000 average first-year contract, and a £6,000 a month engagement covering paid search, landing pages and tracking.
Spend to month six: 6 × £6,000 = £36,000 Enquiries by month six (illustrative): 48, of which 20 qualified Qualified enquiries that became opportunities: 8 Opportunities at proposal stage by month six: 4 Pipeline at proposal stage: 4 × £40,000 = £160,000 Pipeline to spend ratio: £160,000 ÷ £36,000 = 4.4 to 1
At month six that firm has no closed revenue from the engagement and a defensible decision to continue, because stage movement is visible, spend is below created pipeline, and the first proposals land before the cycle's nine months are up. Reverse the numbers, four qualified from 48 and no proposals, and the same spend is a decision to stop, whatever the agency's dashboard says about impressions.
Two points about this example. It is illustrative, with every input chosen for clarity. And the ratio means nothing until you know the firm's proposal-to-close rate; at 25%, that £160,000 of proposals becomes one £40,000 contract, which is why month nine, not month six, is the first fair test of return. Our article on valuing leads when revenue comes months later covers the stage-weighted version of this calculation.
Judge stage movement at six; judge return at nine.
If your website and campaign evidence cannot explain where committee-led deals slow down, map the buying journey with Actualyse — book a call
What tells you nothing about the engagement on its own
Four things get presented as evidence in these engagements and are not.
Rank positions. Five keywords moving from position 40 to 30 is the search engine learning about the pages; it is not a customer. A small business owner on a forum described paying an SEO retainer for three months, watching keywords sit at positions 35 to 45, and seeing no enquiries; the honest reply was that the fee could not fund the work and that "SEO takes time" was how the retainer was being kept alive. Both parties were measuring positions because neither had agreed what else to measure. Google's own documentation on core updates says some changes take days to show and others several months, with the next update sometimes needed before an effect appears; that is a reason to agree leading indicators other than rank, not a reason to wait in silence.
Impressions and reach. Reach is the input to demand creation and it is real work. Reported alone, it is also unfalsifiable. The 95:5 observation reported by Marketing Week in 2021, from Professor John Dawes of the Ehrenberg-Bass Institute and the LinkedIn B2B Institute, holds that only about 5% of business buyers are in market at any one time; reach among the other 95% is the point of brand work, and it will not show as enquiries this quarter. Agree that in advance, or impressions become the excuse.
A dashboard. A dashboard is a presentation layer. It shows what it was built to show. The question is whether the figures on it can be reproduced from your CRM and your bank statements.
A one-month test. It tests setup speed. On a long cycle it tests nothing about selling.
The honest limit of the three-checkpoint approach is that it assumes a reasonably stable market. A firm whose sector stopped buying in month four will see every leading indicator fall for reasons no engagement caused, and the kill criteria will fire on the wrong culprit. When that happens, the month-six question changes from "is it working" to "is anyone buying", and the answer comes from sales, not marketing. The pillar on why B2B sales cycles are getting longer covers that wider shift.
The falsifiable claim here: on a sales cycle of six months or longer, no channel can show closed revenue caused by the engagement's own work inside one cycle length. If an engagement shows you closed-won revenue at month three on a nine-month cycle, either the cycle is shorter than you think, or the revenue was already in motion before the engagement started.
Positions, impressions and dashboards describe effort; stage movement describes selling.
What the engagement contract should say
Most disputes at the end of an engagement trace back to a contract that specified deliverables and never specified evidence. A firm that collects debts from marketing clients described hearing "I'm not paying because the marketing didn't work" as a routine dispute; that sentence is only possible when nobody wrote down what working would look like at month one, three and six.
Put five things in writing:
- The checkpoints. Month one reports tracking, schedule and quality. Month three reports growth, scaled spend and named winners. Month six reports losers, new winners and pipeline by stage against spend. Each with a date.
- The kill criteria. The three above, or your own, stated as a trigger and an action. If the trigger fires, either party can end the engagement at the next checkpoint without penalty.
- Access. Read access for the firm to every ad account, analytics property and tracking container from day one. Reports you cannot check are opinions.
- What the firm keeps. Accounts, data, pages, the tracking build and the account list stay with the firm on exit.
- The first fair test of return. One full sales cycle from the first qualified enquiry. Both parties agree not to claim or dispute return before that date.
If you are about to sign, the growth marketing service page sets out how Actualyse structures an engagement around these checkpoints, and the case studies show the kinds of firms it has suited. If you are three months in with no month-one report, ask for it this week. Its absence is itself the finding.
Write down what working looks like, by month, before the first invoice.
FAQ
Is three months too early to judge a marketing agency?
Too early for outcomes on any cycle longer than three months, and too late for inputs. By month three you should already have checked tracking, schedule and lead quality, and the month-three report should name one to four winners. No winners and no hypothesis is the trigger to act.
How long should a B2B marketing trial run?
Long enough to see stage movement, which on a nine-month cycle is six months, and long enough to see return, which is one full cycle from the first qualified enquiry. A one-month test only tests setup.
What should month one of an engagement report?
Tracking working end to end into the CRM, the first ninety days' schedule and progress against it, and the share of enquiries that match the ideal customer profile. Nothing about revenue.
What is a reasonable pipeline-to-spend ratio at month six?
Created qualified pipeline above spend to date, with the trend from month three rising. Below spend and flat is the kill trigger. The ratio only becomes a return once your proposal-to-close rate is applied, so treat it as a progress measure.
What if the agency will not give account access?
Treat every report as unverified until it does, and make access a condition of continuing. Read access costs the agency nothing and tells you whether the figures can be reproduced.
Summary
- Judge inputs at month one: tracking end to end, schedule, lead quality. Kill if tracking is not fixed by day 45.
- Judge learning at month three: growth in qualified enquiries, spend scaled towards one to four named winners. Kill if there is no winner and no hypothesis.
- Judge compounding at month six: losers stopped, new winners found, qualified pipeline by stage above spend. Kill if pipeline is below spend and the trend is flat.
- Judge return only after one full sales cycle from the first qualified enquiry, and write that date into the contract.
- Require read access to every account from day one, and treat any figure you cannot reproduce as an opinion.
Actualyse builds websites and campaigns designed for long, committee-driven B2B sales cycles. Book a call to talk through where yours stands.
