How to Measure Content Marketing ROI as a B2B Founder

Impressions, likes, and follower growth are the easiest numbers to track and the least connected to revenue. They go up when content is loud, not necessarily when it's working — a post can get thousands of impressions from people who will never buy anything from you, and a quiet 40-view article can bring in a customer worth six figures. Ask most founders what their content marketing ROI is and you'll get a shrug, or a vanity-metrics screenshot that doesn't actually answer the question.

Neither response is a personal failing. Content marketing ROI is a genuinely harder thing to measure than paid ad ROI, and pretending otherwise — by importing a paid-media dashboard and expecting it to explain content — is where most of the confusion starts.

Why content ROI is genuinely hard to measure

Paid ads have a clean trail: click an ad, land on a page, convert or don't, all inside one platform's tracking pixel. Content doesn't work that way. Someone reads an article on a Tuesday, forgets about it, sees your name mentioned in a Slack community three weeks later, Googles you directly, and books a call. Nothing in that chain fires a UTM parameter. The touch that mattered most — the article — is invisible to any tool that only counts last-click conversions.

This isn't a tooling problem you can buy your way out of at seed stage. Multi-touch attribution software assumes enough volume for patterns to be statistically meaningful, and enough of a data budget to stitch web analytics to your CRM to your sales calls. A two- or three-person team has neither, and the software costs more per month than the insight is worth at that scale.

There's also a cost to getting this wrong in the other direction — trusting a last-click report that credits the final Google search someone did, and cutting the content program that actually created the demand behind that search two months earlier.

So the goal isn't perfect attribution. It's a short list of numbers that are directionally honest, cheap to collect, and checked regularly against what your sales conversations are actually telling you.

The six numbers worth tracking monthly

• Inbound conversations started. Replies, DMs, comments that turn into a real conversation, and inbound demo requests that mention a specific piece of content by name. This is the earliest, most direct signal that content is doing its job.
• Pipeline sourced or influenced. Ask every new lead one question — "how did you hear about us?" — and log the answer. Even a rough, self-reported tally beats no attribution at all, and it's usually accurate enough to guide decisions.
• Search rankings for the questions your buyers ask. Track your position for the five to ten specific questions your actual buyers search, not generic industry keywords with volume but no buying intent behind them. Movement here is a leading indicator of pipeline that shows up one to two quarters later.
• Sales-cycle friction, reported by your sales team. Ask whoever runs sales calls whether prospects arrive already familiar with your positioning, or need it explained from scratch. A shift toward "they already got it" is content working, even with no dashboard number attached.
• What sales proactively forwards mid-deal. When a rep sends a prospect a specific article to answer an objection instead of writing a paragraph from scratch, that's content doing real work inside a live deal — a much stronger signal than any view count.
• Return visits from the same account. If your analytics can surface company names, watch for one account showing up more than once across different pieces. Multiple people at one target company reading multiple articles is what a buying committee doing its homework looks like, well before anyone fills out a form.

Three ways the number gets inflated or hidden

Three mistakes distort this more reliably than any gap in tooling does.

The first is treating "influenced" and "sourced" as interchangeable. A prospect who skimmed one article before ever hearing your name isn't the same as a prospect whose entire evaluation traces back to a specific piece of content. Collapse the two together and every deal in the pipeline ends up credited to content, whether it earned that credit or not — inflating the number in a way that eventually gets caught and discredits the whole tracking system.

The second is crediting branded search. Someone typing your company name into Google didn't discover you through content — they already knew who you were and went looking. That's a warm lead following up, not evidence an article created demand out of nothing. Keep branded search out of the content-sourced column entirely, even though it's tempting to count since it's the easiest number to pull.

The third mistake runs the opposite direction: distrusting self-reported data because it feels unrigorous, and quietly reverting to no attribution at all rather than living with an imperfect one. A rough "how did you hear about us?" answer, logged on every call without exception, beats a more defensible-looking system that nobody actually uses. Perfect measurement you don't do measures nothing.

Why month one always looks like a loss

Content is a compounding asset, not a campaign — an article published this month keeps earning search traffic and inbound interest for years, while a paid ad stops the moment you stop paying for it. That means the metrics above will look unremarkable in month one and two almost regardless of quality, because compounding assets need time to accumulate. Founders who judge content ROI on a 30-day view are, structurally, judging it before it's had time to exist.

A practical benchmark: judge content ROI on a rolling 90-day view, not a monthly one, and expect the curve to bend upward starting around month three as earlier content keeps working in the background while new content keeps publishing on top of it.

This is also why cutting a program at day 45 because "it's not working yet" is usually a mistake made from an incomplete read of the data, not a correct read of a genuinely bad program. The fix isn't blind patience — zero inbound conversations and zero ranking movement by month three is a real warning sign — it's judging the program on the timescale content actually operates on.

It helps to separate leading indicators from lagging ones before you start watching the numbers move. Search rankings, inbound conversations, and account-level return visits are leading — they shift first. Pipeline sourced and closed revenue are lagging — they're the point of the exercise, but they arrive last, after a leading indicator has already moved and a sales cycle has run its course on top of that.

What it means when the numbers don't move

Past the 90-day mark, if the six numbers above are still flat, that's real information — but it's easy to misdiagnose. Three patterns account for most stalled content programs:

• Nobody's reading it. Search rankings aren't moving and there's no meaningful traffic. Usually a distribution problem, not a writing problem — the content exists somewhere buyers never look. Fix: publish where your specific buyers already spend time, not more of the same thing into the same silence.
• People are reading it, nobody's talking. Traffic and rankings are fine, but inbound conversations and sales-team feedback stay at zero. Usually means the content answers questions adjacent to your buyer's problem instead of the problem itself — informative, but not specific enough to prompt a reply.
• People are reading and replying, deals aren't closing. Inbound conversations are up, but pipeline and sales-cycle friction don't move. Rarely a content problem — usually a sign the content is attracting the wrong audience, or that something downstream (pricing, product fit, sales follow-up) is the real bottleneck.

Each calls for a different fix, and none of them is "write more content and hope." Knowing which one you're in is the entire point of tracking these numbers in the first place.

A worked example: one quarter, walked through

Say a two-person founding team starts publishing two articles a week in January, aimed at the questions their actual buyers ask before a demo. Weeks one and two: nothing changes — no new inbound conversations, no ranking movement. If they were watching only closed revenue, this would already look like a failure.

By week six, one article starts ranking for a specific buyer question, and a prospect mentions it unprompted on a discovery call — an inbound conversation, plus a self-reported source in the spreadsheet. By week ten, more articles have started to rank, sales reports two prospects arrived already understanding the product's positioning, and a rep forwards an article mid-deal to answer a technical objection. None of this is revenue yet — those deals won't close for six to eight weeks — but every leading indicator has moved.

By week thirteen, the first content-sourced deal closes, tracing back to the article that ranked in week six. Judged against a monthly revenue target from week one, this quarter looked like a near-total loss for ten weeks straight. Judged against the six numbers on a 90-day view, it looks like exactly what working content marketing looks like.

Putting a number on it: cost against return

The six numbers above answer whether content is working. They don't answer the question "ROI" technically asks: what did this cost, against what it returned? Most founders skip that half because the cost side is inconvenient to write down — a freelancer invoice, a founder's own unpaid hours, a couple of lightweight tools — and inconvenient numbers have a way of staying out of the spreadsheet.

Go back to the worked example above. Say that two-person team paid a freelance writer $3,500 a month, plus $150 a month across a couple of lightweight tools — call it $10,950 for the quarter, all in. The deal that closed in week thirteen was worth $22,000 in first-year contract value. Self-reported source data also ties two more open deals, worth a combined $31,000, back to the same handful of articles. Counted conservatively — closed revenue only, nothing still in the pipeline — that's roughly a 2x return for the quarter. Counted against pipeline sourced as well, it's closer to 5x, though that half of the number is soft until those deals actually close one way or the other.

Neither figure needs to be precise to be useful. What matters is the order of magnitude and the direction it's moving in. A program running at 3x–5x and trending upward by month six is worth funding harder, not questioning. A program still sitting under 1x after two full quarters, with the leading indicators from earlier in this guide flat as well, is the real signal to cut it — not one slow month, and not a general feeling that content marketing doesn't work for a company like yours.

A tracking system that doesn't need a marketing stack

You don't need HubSpot, a data warehouse, or a UTM taxonomy to run this well. A shared spreadsheet with a handful of columns does the job for most pre-seed and seed-stage teams:

• Lead source, self-reported. One column, filled in from the "how did you hear about us?" answer, on every single new contact — no exceptions, because the gaps are where the real signal leaks out. Ask it live, rather than burying it in a signup form nobody reads carefully.
• Which piece, if known. Not required, but if a prospect names a specific article, write it down. Over a few months you'll see which pieces keep getting mentioned by name — a better content roadmap than any keyword tool.
• Deal stage and outcome. So "sourced by content" can eventually be checked against "closed" rather than just "arrived."

Tracking the search-ranking number from earlier doesn't require a paid rank-tracker either. Google Search Console is free, already has your data once your site is verified, and shows the exact queries people used to find each page, your average position for each one, and click-through rate — a closer read on real buyer intent than a generic keyword-volume number with no faces behind it.

When a prospect genuinely can't remember how they found you, resist the urge to leave the field blank. Log it as unprompted or organic — it still tells you something different from a lead who arrived through a warm intro or a paid channel.

Pair the spreadsheet with a monthly — not weekly — thirty-minute review: pull the six numbers above, glance at the spreadsheet, and ask whoever runs sales what they're hearing on calls. Weekly reviews mostly just measure noise, for the same reason a 30-day view of a compounding asset looks flat. Monthly is frequent enough to catch a real problem and infrequent enough to actually see a trend.

Where FirstOrg fits in

FirstOrg keeps publishing on the cadence that makes the compounding curve above actually happen, and surfaces the inbound activity — replies, mentions, search movement — tied back to what was published, so you're looking at the six numbers that matter instead of a wall of impressions. It doesn't replace the sales-side half of this system: the "how did you hear about us?" question and the monthly review are still yours to run. What it removes is the excuse that there was never enough content published on a regular enough cadence for the numbers to mean anything in the first place.