A 14-person export company published two posts a month for a year. Twenty-four posts, most of them company news and translated product copy. Then the owner asked one question in a planning meeting: which of those 24 brought us inquiries?

Nobody could answer. The inquiries sat in a shared inbox, the posts sat in a CMS, and the link between the two was never built. So marketing asked for more content, sales asked for better targeting, and next year's plan looked like last year's. *(A typical scenario, composite: patterns we see across export teams, not one named company.)*

If you are the marketing or sales lead at a 5 to 50 person export business, that is your situation. You do not need a bigger content budget. You need the smallest system that answers one question: which piece of content produced this inquiry?

Here is the version we hand teams like this. We are not going to claim we ran this exact playbook at one company and have the numbers to show, because we would rather hand you a method than a story. What follows is the plan we set up, the parts that commonly break, and what it costs in hours.

The 90-day plan

We do not start with a channel plan. We start with the inquiry path, because a content system with no path to an inquiry is a publishing habit.

Two rules sit behind the schedule below. First, one content type until it works. A team that runs guides, case studies and social posts in month one has three chances to fail and no signal about which one failed. Second, the path from a page to a reply gets built in week 1, not week 9. Building it late is the most common way this project dies quietly.

| Weeks | What you produce | Who writes it | Where it goes | What "done" looks like | |---|---|---|---|---| | 1 to 2 | One flagship guide (the page a buyer would ask for), two landing pages with UTMs, one form | Content owner writes, one salesperson answers 30 minutes of buyer questions, boss approves the topic list | Your own site | A real inquiry can be traced back to a page | | 3 to 4 | 2 reply posts a week, each answering one question from the inquiry inbox | Whoever takes the calls, 30 minutes per question, a writer turns it into a post | Blog plus LinkedIn | Sales has forwarded one of them to a live customer | | 5 to 6 | 1 buying guide a week, plus one FAQ page | Content owner, with a salesperson who knows the spec sheet | Blog | Each guide has a table of specs a buyer can copy into an RFQ | | 7 to 8 | 2 anonymized case write-ups a week, with the decision inside them | Whoever ran the project, 45 minutes each | Blog plus email newsletter | The newest inquiry references something that only appears in these posts | | 9 to 10 | Re-cut the 4 best posts into 8 LinkedIn posts and one short video | Content owner | LinkedIn, email, industry groups | Some inquiries arrive from a path other than branded search | | 11 to 12 | 2 evergreen posts a week, plus a monthly attribution review | Content owner plus one salesperson | Blog, email | You can name the two posts that produce inquiries and the three that produce nothing |

Two notes. The rep interviews in weeks 3 to 6 are the cheapest input you will ever get and the easiest to skip, because they interrupt selling. Block the 30 minutes. And the last two weeks are a deletion meeting. If a piece has not earned a reply in three months, it goes on a "maybe later" list.

Where the topics come from

Not from a keyword tool. Keyword tools tell you what already ranks, and what already ranks is the output of people who have done this for ten years. The topics that bring inquiries come from four places you already own.

1. The inquiry inbox, last 90 days

Export the emails. One column for the question the buyer was actually asking, not the one they typed. "How long does it take to ship?" means "will you be late on my schedule, and who do I blame when it happens." Twenty emails will give you eight topics, because the same question keeps arriving in different words. *(A typical scenario, composite: in inbound mail a handful of questions repeating is the norm.)*

2. What buyers asked on the call

Ask each rep for the three questions buyers asked most last month. Reps undercount, and in a useful direction: they forget the question they answered forty times. Give them a shared sheet with three columns (question, how often, who asked) and that sheet is your first editorial calendar.

3. The quote template and the spec sheet

Every blank line in the quotation is a topic. "Lead time: ___", "Certification: ___", "Payment terms: ___". Buyers fill those blanks with questions, and each one is a page somebody can copy into an RFQ. The fastest source you have, sitting in a file right now.

4. The evaluation checklist from a deal you lost

Ask sales for the checklist a buyer's procurement team used. When you lose, somebody usually has a copy of what they scored you on. A lost deal is a free content brief, and most teams throw it away. If there is no checklist, ask the buyer one question: what would you have needed from us? Write the answer down verbatim.

Each source costs about 30 minutes. Run them once and you have enough for two months.

How to tell if a piece of content is worth publishing

Before anything ships, we run four questions. Not a score, not a KPI. Four yes or no questions, and the piece needs three of the four. The paste test. Can a buyer copy a table or a paragraph out of this and paste it into an RFQ or an email to their boss without editing it first? If not, they will not.

The knowledge delta. Does the piece contain at least one number, specification, threshold or decision rule that the buyer did not have before reading it? A restatement of your own homepage fails this, every time.

The forward test. Can a rep send this to a live inquiry with one line of context, no rewriting? If it needs a rewrite it will not get forwarded, and a piece nobody forwards is a piece nobody reads.

The hallway test. Could you answer this yourself in sixty seconds if somebody asked you in the corridor? If you cannot, you are not ready. You will write something vague to cover the gap, and vague content attracts nobody.

Now the honest part. This test tells you whether a piece deserves to exist. It does not tell you how it will rank, and we will not promise inquiry numbers, because we do not have a case that would let us promise them honestly.

Two ways this breaks, and how to fix each

Failure mode 1: you write about you

Symptom: your ten best posts are about your company, your factory, your awards, your capabilities. Your buyer does not care about your factory. They care about whether their problem goes away.

Fix: run a 3 to 1 rule. For every post about you, publish three that answer a buyer's question. Then rewrite one page on your site in the buyer's language and read it back. Usually it does not read like something you would want to read. That rewrite is often the best post of the month.

Failure mode 2: content and sales live in different systems

Symptom: a rep asks how the buyer heard about us, the buyer says Google, and nobody finds out which page. The content is anonymous, so it gets no credit, so it gets no budget next year. This is the failure that makes a content program look dead when it is only unattributable.

Fix: minimum viable attribution, four parts. One form per campaign with UTMs. One source field on the contact record. A 20 minute weekly review of where the week's inquiries came from. And every inquiry gets a reply within 24 hours, with the piece they read attached if you can find it.

A third break: one person writes everything. The topic list lives in their head, they take two weeks off, the program stops. Keep the list in a shared sheet. It takes an hour to set up and it is the difference between a program and a hobby.

What this actually costs

Our estimate, and it is an estimate: 8 to 12 hours a week for the first quarter, for a team of one to three people touching content. Roughly 30 minutes a day collecting questions, 2 to 3 hours a week writing, 1 hour on distribution, 1 hour on review and cutting. That is about a quarter of one person's time. One person can run it, but not while carrying a full sales quota. If you have to choose between a content hire and a second salesperson right now, take the salesperson.

Tooling can stay cheap. A CMS, a form tool and a spreadsheet cover the first 90 days. If you want content and CRM to talk to each other, that is a platform decision, and SiliMind pricing is on the official site, because deployment model and seat count change the quote and we would rather you look than trust a table in a blog post.

Who should skip this

Four kinds of teams should not start this now.

No stable product line. If you still change what you make every quarter, there is no body of knowledge to write from and every post ages badly. Fix the catalog first.

A boss who never touches it. Content judgment needs somebody who can say no to a bad topic. If the founder's only involvement is forwarding links, nobody can decide what is worth publishing.

Nobody who talks to buyers. If the only inputs come from the product team, you will publish specifications nobody asked for, because the product team is the only room where specifications get discussed.

Fewer than two people with steady weekly time. Below that, outbound is cheaper. One rep sending 50 researched emails a week will often beat a blog nobody reads.

One more: buyers who do not research. In some categories the decision happens in twenty minutes at a trade show, and a content library is the wrong asset. Check how your last five customers bought before you build anything.

Where a platform earns its keep

Content only pays back if you can connect it to the inquiry, and that connection is the genuinely hard part. That is where SiliMind is built to help. On the content side, topic suggestions are scored by search volume, competition difficulty and opportunity, with a weekly topic report, so the shared sheet does not depend on one person's instinct. The content workbench lets a draft quote fragments from documents you have already uploaded (PDF, Word, Markdown, a URL) and cite them in the body, which is how you stop an AI draft from inventing a tolerance your factory cannot hold. On the CRM side, when a contact is created, four attribution rules run (email domain, company name, contact name, country), each weighted, and when the total clears the threshold the matching pieces attach to that contact. Content is also matched to open deals by company name, industry and keyword overlap, so a rep can see which pieces a live deal has already seen. Inquiry and sales conversation then sit on one timeline across 10 channel types, rather than two tabs and a spreadsheet.

We wrote up the mechanics in how content attribution works for B2B marketing teams. If you are weighing tooling, SiliMind vs HubSpot covers where a bundle of separate tools and one integrated platform diverge.

Start tomorrow

Export the last 90 days of inquiry emails. Tag each one with the question the buyer was really asking. Pick the question that came up most often. Publish the answer this week, with a form and a UTM on the page that answers it.

That is the whole first step. The 90 days above come after it.

Appendix (excluded from word count)