+657 net new followers across May, a 4.7% gain. The curve steepened in the final week as the CLOC and Word add-on content compounded.
Two posts (Legal OS is a lie, the off-topic COO job share) drove about 43% of total monthly reach.
Each list below ranks by a different signal: impressions for raw reach, saves for "I want to come back to this" intent, and engagement rate for resonance on smaller-reach posts. The split matters this month: the biggest-reach post was off-topic, while the biggest-intent posts were a framework and a hot take.
Who got seen the most.
B2B's strongest signal. Readers bookmarking to revisit.
Smaller-reach posts that overperformed for their audience.
Text long-form led volume and carried the heavy hitters. Image was a close second.
Text averaged highest, but that average is skewed by two outliers (Legal OS, COO share). Video lagged badly.
We read each of the 16 posts and classified them by the job the post was doing. Six recurring patterns emerged. The breakdown below shows volume and performance per pattern so we can see which deserve more calendar share next month and which to retire.
Off-topic tops the chart on a single viral post. Among on-brand patterns, hot takes have by far the highest ceiling.
Saves are the strongest intent signal. Hot takes and frameworks own this metric.
Six content patterns observed across 16 published posts.
| Pattern | Posts | Total imp. | Avg imp./post | Total saves | Top performer |
|---|---|---|---|---|---|
| Hot takes vs incumbents | 5 | 60,075 | 12,015 | 129 | Legal OS is a lie (30.7K imp) |
| Frameworks | 2 | 11,010 | 5,505 | 85 | 5 to 7 stack (68 saves) |
| Founder arc / CLOC journey | 2 | 17,830 | 8,915 | 7 | End of CLOC (11.0K imp) |
| Product/company updates | 2 | 7,746 | 3,873 | 7 | Word add-on (5.9K imp) |
| Event promotion | 4 | 7,819 | 1,955 | 4 | Kodex dinner invite (3.2K imp) |
| Off-topic / community share | 1 | 25,896 | 25,896 | 65 | COO-for-artist share (25.9K imp) |
Five hot-take posts averaged 12.0K impressions, drove 60.1K total (46% of monthly reach), and produced 129 of the month's 297 saves. Every one opened with a contrarian position against a named incumbent (Legal OS, Harvey, Legora, CLM string-search, "pretty UI"). This is the workhorse pattern for both reach and intent.
Two framework posts drove 85 saves between them. That is 29% of all monthly saves from just 13% of post volume. They also pulled the highest-quality audience of any on-brand pattern (43% ICP fit, see below). Heavily underweighted relative to their return.
The COO-for-artist job share was the 2nd-highest-reach post of the month at 25.9K and pulled 65 saves, but its audience was music, creative, and people-ops, not Legal Ops. ICP fit was 13%. Strong vanity reach, near-zero buyer relevance. One post, treated as an anomaly throughout this report.
4 event-promo posts averaged only 2.0K impressions and drove 4 saves total. They occupied 25% of the calendar and returned 6% of the reach. All three bottom-performing posts of the month were event teases.
The two product posts (Word add-on, localization) averaged 3.9K impressions and pulled the 2nd-highest ICP fit of any pattern (41%), but together generated only 7 saves. These posts attract the right people. They just do not get bookmarked. Pair them with something saveable.
Jenn's ICP for Contracts.AI is in-house Legal Ops and Counsel at corporates: GCs, VPs of Legal, Heads of Legal Operations, Directors of CLM. We sampled lead profiles across 10 posts (covering all 6 content patterns, 154 profiles total) and classified each one by current role, company, and industry. The stratified sample lets us compare ICP fit between content patterns, not just between top performers.
% of engagers who match the buyer profile, by what type of post pulled them in. Higher = better audience targeting per pattern.
Across all 154 sampled profiles. Establishes the baseline for tracking month over month.
Where the in-house Legal Ops + Counsel folks who engaged actually work. Confirms we're hitting the natural buying segments for Contracts.AI.
Director/VP/Head/Chief-level Legal Ops or Counsel folks who engaged. These are buyer-fit accounts worth a closer look from sales.
Aggregate baseline: ~33% of engagers across the stratified sample are core ICP (in-house Legal Ops + Counsel). This becomes the metric we track month over month. Target for June: 38%+.
Frameworks pull the highest ICP fit of any on-brand pattern (43%). They are the strongest content pattern for buyer-targeting AND drive the most saves per post. This is the single strongest argument for shipping more frameworks in June.
Product posts (41%) and the qualified Kodex dinner invite (40%) sit just behind. Product announcements attract practitioners actively evaluating tools. The event ICP figure comes from the curated, CLOC-only dinner list, so treat it as the optimistic end for events. The broad event teases pulled far less.
Hot takes have a lower ICP fit (29%) even though they get the most reach. Why: competitive legal-tech vendors (iManage, Agiloft, Brightflag, Consilio, Lawtrades), consultants, and law firms crash these conversations. The industry is watching, but vendor engagement does not convert. They are still the workhorse for reach. Just do not mistake the volume for ICP signal.
Industry mix maps cleanly to Contracts.AI's natural buying segments: enterprise software (34%), healthcare/pharma (11%), manufacturing/industrial (11%), financial services (9%), retail/consumer (8%), entertainment (6%). No notable misses.
Note on methodology: this is a clean May 1 to 31 calendar-month sample (n=154), distinct from the prior rolling-window cycle, so 33% is the new calendar-month baseline rather than a drop. Event promo (n=15) and off-topic (n=15) are single-post samples and are flagged accordingly. The off-topic COO share is shown for completeness at 13% ICP fit.
The first sentence does almost all of the work on LinkedIn. The algorithm decides reach within the first 90 minutes based on initial dwell and engagement, so the hook is where the post is won or lost. We analyzed the opening lines of the 8 highest-performing posts of the month. Five ingredients appear in nearly every one.
Observed across the 8 top-performing posts (top 5 by impressions and top 5 by saves; overlap counted once).
"15 years", "1,200 contract types", "7-figure CLM", "$600M-funded company", "cloned in 2 weeks". Every on-brand top performer drops a hard number before sentence 3. Numbers create immediate credibility and stop the scroll.
Netflix, Spotify, Cisco as credentials. Microsoft, Anthropic, Harvey, Legora, Ironclad as targets. The winning hooks name names. Generic phrases like "the industry" or "most vendors" do not show up in winning openings.
"the biggest lie in legal tech", "Believing they could was my mistake", "Last year I walked the floor as board president, this year I'm setting up a booth", "the prettier the product, the more skeptical I get". Conflict creates the scroll-stop. Without tension, there is no reason to keep reading.
"Here's what every vendor knows but won't admit", "here it is", "Here's what I think is actually happening", "Here's why". The top hooks set an expectation that the explainer is coming. Without this, readers exit before reaching the substance.
"I've worked", "I fought", "I half expected". The top hooks averaged about 22 words in the first sentence. Personal voice, fast cadence, no warmup paragraph. Corporate "we are excited to share" openings did not appear in any top performer. Note: the one off-brand outlier (COO job share) broke these rules and still hit 25.9K. That was network-share virality on a job posting, not a repeatable hook structure.
The bottom 3 posts (about 19% of the calendar) ranged from 1,186 to 1,740 impressions. All three were event promos. Each missed multiple hook ingredients or hit a theme conflict. Hypotheses below.
Opens with "Most teams going to LegalTechTalk are still manually stitching together contract data..." which has the shape of a hook but contains no number, no named target, no tension. Then it pivots straight to "we'll be in London, stop by our table." Missing 4 of the 5 hook ingredients, posted at 8am ET before the audience logs in, and in video format which underperformed all month. The classic event-tease failure: no payoff for reading past the first line.
This was the third CLOC-dinner logistics post in a five-day window (after the May 1 "come to dinner" text and alongside the May 6 Booth 400 tease). The feed had already seen the invite, so the video added no new information and the algorithm had little reason to expand it. Pure RSVP, no thesis, no hook. When several near-identical event posts cluster, only the first earns reach.
"Excited to be heading to Chicago... at Booth 400... hosting a dinner Tuesday night." It is an announcement, not an argument. No number in the hook, no tension, no payoff promise, and it landed in the same CLOC-promo cluster as two other posts. Decent 3.6% engagement rate from a small, warm audience, but almost no reach because there was nothing to expand.
1. Clustering several event-logistics posts within a few days of each other (theme cannibalization).
2. Event teases without a hook (number + named target + tension + payoff).
3. Publishing before 9am ET, where every underperformer this month landed.
4. Chasing off-topic reach: it inflates impressions but pulls the wrong audience and trains the feed away from Legal Ops.
Six concrete moves, each grounded in something we observed this month.
Posts that open with a contrarian position against a named incumbent are this account's most reliable winner for both reach and saves. Aim for 2 per week, each framed against a specific named entity (Ironclad, Harvey, Legora, Microsoft's Word-native drafting, the "Legal OS" narrative, CLM renewal pricing). Sequence them at least 4 days apart to avoid the cannibalization we saw on May 21 to 22.
Frameworks are massively underweighted relative to their return. Two posts drove 29% of monthly saves AND hit 43% ICP fit (highest of any on-brand pattern). June should ship roughly one framework per week. Candidates: "the 4 questions a Legal Ops buyer should ask any AI vendor," "post-signature workflow in 5 phases," "what an enterprise-ready data schema looks like." Each as an image carousel.
The 4 event teases averaged 2.0K impressions and 1 save each, and were the entire bottom of the calendar. The replacement: after every conference, publish a "5 things buyers told me on the floor" or "what changed in the room vs last year" post. LegalTechTalk London is mid-June, so ship a thesis recap, not a "stop by our table" tease. Cap pure event-tease posts at 1 per conference.
The month's biggest posts (Legal OS, 5-to-7 stack, MikeOSS, the CLOC recaps) all published between roughly 10:30 and 11:00am ET on weekdays. Every underperformer published at 8am ET, before the audience logged in. Stop posting before 9am ET, skip weekends, and keep hot takes at least 4 days apart to avoid theme cannibalization (Legal OS on May 21 then State of legal tech marketing on May 22 cost the second post about 9x the reach).
Product posts pulled the 2nd-highest ICP fit of the month (41%) but almost no saves: Word add-on (5.9K imp, 6 saves) and localization (1.8K imp, 1 save). They bring the right buyers but give them nothing to bookmark. Pair each June product announcement with a "how to evaluate this" framework within 72 hours so the buyer attention converts into intent.
The COO-for-artist job share was the 2nd-biggest post of the month at 25.9K impressions, but it pulled a music, creative, and people-ops audience (13% ICP fit) and has no pipeline relevance. High vanity reach trains the feed away from Legal Ops. If Jenn wants to do a network favor like this, fine, but do not count it as performance and do not let it set the content pattern.