The phrase “fractional CMO” has been stretched to cover a strategist who sends you a deck, a consultant who joins your Monday meeting, and a contractor who runs your ads. Those are three different jobs at three different prices, and the confusion is expensive for buyers.
So here is what the first ninety days looks like when it is done properly, in enough detail that you can hold a conversation about it. If you are interviewing someone for this and their answer to “what happens in month one” is a list of deliverables rather than a list of questions, keep interviewing.
One thing to establish up front. A fractional CMO is not a cheaper full-time CMO. It is a different shape. You get senior judgment applied to a small number of decisions, plus enough hands to make sure the decisions turn into work. What you do not get is somebody in every meeting, and if what you need is somebody in every meeting, hire.
Days 1 to 14: find out what is actually true
The first two weeks are almost entirely diagnostic, and a good operator will resist producing anything that looks like a strategy during them.
The numbers, from the source. Not the dashboard, the source. Revenue by month for at least twenty-four months, split by product or service line. Enquiries by month, by channel, and critically by what happened to them. Close rate. Average deal size. Sales cycle length. If those numbers do not exist, that is finding number one and it changes the whole plan, because you cannot allocate a budget against outcomes nobody is recording.
The analytics, checked rather than read. Almost every GA4 property I inherit has at least one of: conversion events firing on page load, a form thank-you page counting as a conversion when the form itself errors, internal traffic uncounted, or cross-domain tracking broken so that the checkout looks like a referral from your own site. The reason to check before reading is that a plan built on wrong numbers is worse than a plan built on no numbers, because you will defend it.
The sales conversations. Three to five recorded calls, or three to five hours with whoever sells. This is where positioning actually lives. The language customers use to describe their problem is almost never the language on the website, and the gap between the two is usually the single largest available improvement.
The customers themselves. Five interviews if you can get them. Why they bought, what they nearly did instead, what almost stopped them. Twenty minutes each. Nothing in marketing has a better ratio of insight to effort, and almost nobody does it.
The infrastructure. What the website is built on, who can change it and how fast, where the CRM ends and the spreadsheets begin, what is stitched together with somebody’s personal Zapier account. This matters because it determines what is actually possible in the next sixty days, and half of all marketing plans fail on execution constraints nobody checked.
By the end of week two the output is not a strategy. It is a short document listing what is true, what is broken, and what nobody can currently answer. Delivered verbally first, because some of it will be uncomfortable and email is a bad medium for that.
Days 15 to 30: decide, and write it down
Now the judgment gets applied. This is the part you are actually paying for.
The plan has to answer four questions, in order, and each answer has to close off options rather than keep them open.
Who is this for, specifically. Not a persona document. A description precise enough that you can say who it is not for. Most small companies are trying to sell to everyone and are therefore compelling to nobody. Narrowing is the highest-leverage move available and the one that generates the most internal resistance, because it feels like turning down money.
What are we claiming. One sentence, in customer language, that a competitor could not plausibly say. If your competitors could put it on their homepage without anyone noticing, it is not positioning, it is category description.
Where do we compete for attention. Two channels. Possibly three. A small business running six channels badly beats itself; the same budget on two channels done properly is a different company. This is where a good fractional CMO earns their money, because choosing means killing something somebody in the building is emotionally invested in.
What are we going to stop. There is always something. The trade show that produces business cards and no pipeline. The blog nobody reads. The agency retainer that has been quietly renewing since 2023. Finding the stop list is often how the plan pays for itself in the first quarter.
The plan itself should be short. Ten pages is generous. It needs a sequence, a cost against each item, a named owner, and a small number of measures that will tell you within ninety days whether it is working. If it needs a presentation to be understood, it is not finished.
Then it gets presented to whoever holds the budget, argued about, and changed. The argument is part of the process, not a sign it went wrong.
Days 31 to 60: build the boring things first
Month two is where the difference between a strategist and an operator becomes obvious. A strategist hands over the plan and asks how it is going. An operator starts building, and starts with the plumbing.
Measurement, properly. Conversion events that fire on real outcomes. Offline conversions imported where the real close happens in a CRM. A single weekly view of enquiries by source and by quality, not just volume, because volume without quality is how a marketing function ends up hated by sales.
The conversion path. Almost always there is one page doing most of the work and it is almost always underbuilt. Fix that before buying more traffic. Sending more people to a page that does not convert is the most common way to waste a quarter.
Response and follow-up. How fast an enquiry gets answered, and what happens if it is not. This is nominally sales, not marketing, and it moves the number more than most campaigns. If you have to fight about ownership here, fight about it.
Technical foundations if search is a channel. Crawlability, speed, structure. Unglamorous, slow to show results, and pointless to start in month five if it needed starting in month two.
There will be quick wins in this window and a good operator takes them, partly because they are worth taking and partly because a fractional engagement needs visible momentum to survive its first budget review. But the quick wins are not the point of the month. The plumbing is.
Days 61 to 90: run it, read it, adjust it
Month three is the first month where the machine is running and producing readable signal.
The work is now: execute the plan, watch the measures, and change what the evidence says to change. The measures will be noisy, especially at small volumes, and the discipline is to distinguish a trend from a fortnight. Small businesses tend to kill things too early. The counter-discipline is to have decided in month one what “not working” would look like, so that the decision to stop is made against a threshold rather than a mood.
Somewhere in here the second thing gets built. And somewhere in here the first genuinely awkward conversation happens, which is usually one of: the positioning is right but the product does not match it, the channel that works is the one nobody wanted, or the pipeline problem was never a marketing problem and the plan cannot fix it. All three are worth more than a good campaign, and all three are why you brought in someone from outside.
What should exist on day 91
If the ninety days worked, these things are true and you can check every one of them yourself:
- You can say who your best customer is, specifically, and so can your salespeople, in the same words.
- There is a written plan, under ten pages, with owners and costs, that people actually refer to.
- The numbers in your analytics are trustworthy, and somebody can explain what each conversion event means.
- You know your cost per enquiry and, at least roughly, your cost per customer, by channel.
- Something has been stopped.
- At least two substantive things have been built, not just recommended.
- You have a weekly rhythm that survives without the fractional CMO in the room.
That last one is the real test. The job is to install judgment and a system, not to become load-bearing. Any engagement that would collapse the week it ends has failed at the thing it exists to do.
The honest failure modes
Not enough days. Below two days a month there is not enough continuity to be accountable for anything, and the engagement degrades into advice. Most of the ones that go wrong were underscoped from the start, usually because the buyer was testing the water.
Judgment with no hands. A plan handed to a team with no capacity to build it produces nothing and a lot of resentment. Either the fractional role includes execution, or execution is funded separately and explicitly. It cannot be assumed.
No authority. If every decision has to go to a committee that meets monthly, ninety days is one and a half decisions. Someone has to be able to say yes.
The wrong problem. Sometimes the honest answer after two weeks is that marketing is not the constraint. Pricing is, or churn is, or the product is. A good operator tells you that and offers to leave. It is worth asking, on the first call, what they would do if they concluded that, because the answer tells you most of what you need to know about who you are hiring.