A sample marketing strategy for a 30-person B2B company, week by week

Here is a sample marketing strategy for a 30-person B2B company with no marketing team: six weeks, four channels (search, AI answers, LinkedIn and outbound), and one review a week. Most examples are written for a marketing department. This one is written for a company that does not have one: thirty people, a founder who sells, a product that works, and no one whose job is demand. The example company is invented, and every number in it is an illustration, not a measurement. The structure is the part to copy.

The strategy runs six weeks, not a quarter. A first plan should be short enough to review before habits set, and long enough to show whether anything moved. Search results take weeks to register; LinkedIn tells you in days; outbound tells you in a fortnight. Six weeks lets all three report once.

Before week one: three inputs

• Who buys, in one paragraph. Industry, size, the role that signs, the problem that makes them look. Every piece below is judged against it.
• A visibility baseline. Where the site ranks for the questions buyers ask, and whether ChatGPT, Perplexity, Gemini or Google's AI Overviews name you when asked those questions. The number is usually low. Write it down anyway; it is what week six is compared against.
• What you will never say. Claims you cannot back, competitors you will not name, prices you do not publish. Rules written once save an argument every week.

The shape of this B2B marketing strategy example

The plan has three stages that run in parallel, weighted differently each week. Find demand is research and measurement: which questions, which competitors, which gaps. Create is the inbound work: articles for search and AI answers, posts on LinkedIn. Get leads is outbound and capture: the people who fit, reached; the companies that came, named.

For LinkedIn specifically the plan uses a 40/40/20 mix, borrowed from how strong founder accounts are actually run: 40 percent of posts teach something from your work, 40 percent take a position on how the buyer's problem should be solved, and 20 percent point at what you sell. Three posts a week is the floor at which an account grows; the mix keeps it from becoming a sales feed.

Weeks one and two: learn and open

Search. Pick two keyword clusters where buyers search, competitors rank and you do not. Brief one article per cluster: who it is for, the angle, the questions it must answer in its first lines, the sources it may cite. Approve the briefs before a word is drafted; a wrong brief costs a week.

AI answers. Approve ten buyer questions and start the weekly check across the four assistants. Do nothing else with it yet; two weeks of readings are the baseline.

LinkedIn. Three posts each week from the founder: two teaching posts from real work, one position post. No selling in the first fortnight. The aim is to be worth following before you ask for anything.

Outbound. Define the list only: the roles and company shapes that fit, from your own prospect database. Build one sequence and run it as a rehearsal, where every message is written and every step timed but nothing is sent. Read it as the recipient would.

Weeks three and four: publish and reach

Search. The two articles go to the site as drafts; approve and publish. Brief the next two, this time from the gaps the AI visibility check exposed: questions where you were absent and a competitor's comparison page was cited.

AI answers. Apply one fix per gap: an answer-first rewrite of an existing page's opening, or a new short page for a question nobody on your site answers. Check the robots rules while you are there; an accidental block of an AI crawler is common.

LinkedIn. Keep the cadence and add the first 20 percent: one post that shows what the product does, told as a customer's problem rather than a feature list. Look at who engaged. The companies engaging are a lead source most teams never read.

Outbound. Go live with the rehearsed sequence to a small batch. Conditions in the sequence, such as "if they accepted the invite", should be decided when the wait ends, not when the first message is sent. Stop on reply.

Weeks five and six: measure and steer

Read the companies that came. Which visiting companies fit the profile, what they read, and which article or post brought them. This is the first real evidence in the plan. A company that fits, arrived from an article and read pricing is worth more than any impressions figure.

Grade every task as improved, flat, declined or still measuring, and say how the result is attributed. Do not claim causation the data does not show. Search at week six is mostly "still measuring", and that is an honest reading.

Steer, do not restart. Keep what brought fits. Replace one cluster that brought nothing. Adjust the LinkedIn mix only if the position posts are the ones the right companies engage with, which is usually the case. Then extend the plan by another six weeks with the same shape.

From a marketing plan template to your own plan

Mati is an AI that runs your marketing end to end: it builds and runs this plan from your own data rather than from a template. Onboarding takes minutes: it reads the site, the search data and the LinkedIn account, asks the few questions only you can answer, and produces a dated plan with every task explained: why this cluster, why this week, what it should move. Your market and competitors are compared daily and your keyword clusters refreshed weekly, so the picture stays current. The writers draft the articles and posts in your voice, from your knowledge base, in minutes; you approve each one. The AI visibility check runs weekly. Mati names the companies that came, and outbound to the people who fit rehearses before it goes live. A weekly review proposes changes with the evidence attached, and you apply the ones you want. A team of two or three runs what would otherwise take an agency or a stack of tools and the people to run them.

The plan above is a sample. Yours will differ in the clusters, the questions and the mix, because those come from your numbers. The shape, six weeks across three stages with approval at every gate, is the part that holds.