362,542 Impressions in 90 Days: How We Ran LinkedIn for a UK Construction Firm

Between the start of the 90-day window and the end, the founder's LinkedIn account for this UK construction client hit:

• 362,542 impressions — a +1,104% year-on-year increase versus the same 90-day period in the previous year.
• 166,659 unique people reached.
• 1,731 engagements across all reactions, comments, reshares, clicks and follows.
• 1,456 reactions and 199 comments.
• +2,134 new targeted followers, roughly 74% of whom matched our ideal buyer persona by function and seniority.
• 47 inbound conversations — unsolicited DMs and connection-request messages mentioning specific posts.
• 12 qualified opportunities, 4 closed-won contracts, and approximately £380,000 in new ARR attributed to the channel over the window and the quarter following.
• £0 in paid LinkedIn spend. All organic.

This wasn't a fluke — the month immediately before the window had already set up the conditions. But those are the numbers that showed up in the 90-day report, and they're the numbers that made the client's board take the LinkedIn channel as seriously as their trade show budget.

If you opened the analytics at month 8 of the engagement, you would have seen an account that looked fine, not special. Monthly impressions sat around 18-35k. Followers were growing at roughly 120 a month. The founder was posting 3-4 times a week, attending to comments daily, and running 15 connection requests a week into the target personas.

The temptation at that point is always to change the strategy. Most clients — and most founders doing this themselves — would have panicked around month 6 and pivoted: tried short-form video, bought a ghostwriter, run ads, added a second poster. We held the line. The reasons:

• The follower quality was improving every month. By month 8 the audience was roughly 68% operational roles (PMs, commercial, planning) and 12% innovation/digital — the exact split we wanted.
• The comment quality was shifting from generic to specific. People were referencing live projects and asking technical questions. That's the signal the algorithm reads as authority.
• The inbound DM quality had quietly moved from spam to senior. We were getting one or two unsolicited introductions a week from Tier 1 names.

The compounding happens under the surface before it appears in the impression graph. The graph is a lagging indicator.

When we broke down the 90-day window post by post, three content types accounted for the overwhelming majority of reach.

1. Operator-grade site-reality posts (41% of total impressions). Screenshots from live jobs — redacted RFIs, clash detection stills, programme fragments, BoQ extracts. One post of a redacted NCR with a three-paragraph analysis of the root cause hit 74k impressions on its own. What makes these work: they cannot be faked. Every line is proof of on-site expertise.

2. One viral teardown post (22% of total impressions). A single 1,800-word LinkedIn article that deconstructed a published Tier 1 case study and pointed out, civilly, where the marketing had outrun the underlying engineering. It hit 79,000 impressions, 340 reactions, and got reshared by four well-known industry figures. This post alone drove ~150 qualified profile views and 9 inbound DMs.

3. A founder-led hot-take series on a regulatory shift (17% of total impressions). The Building Safety Act secondary legislation landed during the window. The founder published a six-part series over three weeks with specific, non-generic takes on what the clauses meant for mid-market contractors. Each post hit 15-25k impressions; the series collectively drove a measurable step-change in followers and inbound enquiries.

The remaining 20% came from a long tail of medium-performing posts, reshares, and comments on other people's content.

Equally useful: the content we tried and actively cut.

• Short-form video. We produced six polished 45-second videos in month 5. Total impressions across all six: ~11,000. A single site-photo text post outperformed all six combined. Construction LinkedIn is not a video-first channel and we stopped forcing it.
• Founder "journey" posts. Classic SaaS fodder — "three things I learned building X". Consistent underperformance, typically 40-60% of account average. The audience did not want founder-as-protagonist content.
• Industry news reshares without a take. Sharing a Construction News article with "great read" — the algorithm punished us for it and the audience yawned. Once we added a sharp editorial take to every reshare, these became top-quartile performers.
• Daily posting. We ran daily for 6 weeks in month 7. Reach per post dropped by roughly 38% versus the 3-4x-a-week baseline. We cut back and reach recovered inside two weeks.

For the 90-day window, here was the weekly operating rhythm:

• Monday 07:30 — site-reality post (screenshot or photo + 150-200 words).
• Monday-Friday before 10:00 — reply to 10-15 comments on other people's posts in the target audience.
• Tuesday — 15 personalised connection requests. Zero pitches in the request. Follow-up only after engagement.
• Wednesday 08:00 — the anchor post of the week: teardown, carousel, or hot take. 400-900 words or 8-12 carousel slides.
• Thursday — 5 warm DMs to people who engaged with the week's content. Reference what they commented on. Never pitch in the first DM.
• Friday — an editorial reshare of an industry news item with a strong take, or a short reflective post.

Total founder time: approximately 3.5 hours per week, with content drafting handled by our team and final polish by the founder. This is the time budget that compounds.

LinkedIn analytics are interesting. Revenue is what matters. In the 90-day window and the quarter immediately following, the client recorded:

• 47 inbound conversations — DMs or connection-request messages that explicitly referenced specific posts or the company's content.
• 23 demo calls booked directly off those conversations.
• 12 qualified opportunities entered the pipeline with properly-scoped requirements.
• 4 closed-won contracts, with an average ACV of roughly £95k.
• Approximately £380,000 in new ARR attributable to the LinkedIn channel.

To put that in context: the client's trade show spend the year before — Digital Construction Week, Futurebuild, UK Construction Week — totalled around £48,000 and produced 2 closed-won deals worth ~£120k. LinkedIn, at £12,000 in agency retainer across the quarter, produced over 3x the revenue at a quarter of the cost.

That's the ratio that reset the client's board's view of the channel.

Three things in hindsight:

1. Start the anchor-post series earlier. The Wednesday anchor-post discipline kicked in at month 4. If we'd started it in month 1, the compounding would probably have shown up a full quarter earlier.
2. Build a small stable of founder-adjacent voices sooner. By month 9 we had two senior employees posting regularly as well. Their content amplified the founder's — the same reader seeing three different credible voices from the same firm accelerates trust. We should have done that at month 3.
3. Track inbound DMs systematically from day one. We started logging DMs properly in month 7. If we'd done it from day one we would have been able to attribute pipeline to content with more granularity.

So what: The 362,542-impression 90-day window looks like an overnight hockey stick. It wasn't. It was nine months of patient, operator-grade, founder-led content aimed at a clearly-defined construction buyer persona. The window did not invent the audience — it harvested it. The lesson for any ConTech or construction founder reading this: do not change strategy at month 4. Do it properly for nine months and then measure.