The most common cause of death for content programmes at partner firms isn’t bad content. It’s premature evaluation. A programme gets approved, posts start going out, and at month three someone in a partners’ meeting asks how many leads it has produced. The honest answer is none, because ninety days is not how long trust takes to build, and from that moment the marketing manager is defending a programme instead of running one.
The fix happens before launch, not after. Agree what will be measured, and what won’t be, as part of the approval itself. It’s the same expectation-setting discipline that sits behind the whole workload story in the maths doesn’t work: the structure you set up front decides whether you spend your energy building or defending.
Here’s what belongs in the first 90 days, and what doesn’t.
Measure: impressions
LinkedIn counts an impression when a post is on screen for at least three seconds, which makes it a rough but genuine attention signal. Early on, growing impressions tell you the partner’s thinking is reaching more people each week, which is precisely the job at this stage. Track the trend rather than single-post spikes. One post doing unusually well matters far less than the baseline creeping upward month on month.
Measure: posting consistency
This one feels like marking your own homework, and it’s still the most predictive number you have. Nearly every previous content attempt at your firm died of inconsistency: the calendar that went quiet, the newsletter that stopped, the partner who posted twice. A programme that publishes on schedule, for every participating partner, for 90 straight days has already beaten the failure mode that killed everything before it. Consistency is the leading indicator that makes every other result possible, so report it with a straight face.
Measure: changed conversations
The softest measure and the most persuasive one. A prospect who quotes a post in a meeting. A client who raises a topic before you suggest it. A referral that arrives warmer than referrals used to. None of this shows up on a dashboard, so it has to be collected by hand: keep a running log, note the date and context, and ask the partners to pass along anything they hear. By day 90 you’ll have a short list, and one concrete story does more in a partners’ meeting than any chart you could bring.
Don’t measure: leads
Not because leads don’t matter, but because clients don’t move that fast. Engaging a firm changes a client’s business for years, and people make decisions of that size on their own timetable, over weeks, months, sometimes years of quiet reading.
The audience maths makes the same point. On LinkedIn, roughly 1% of people post, about 9% engage with what they read, and 90% watch in complete silence. The people who like and comment are rarely your buyers; the silent 90% is where clients come from, and by definition they leave no trace while they’re deciding. Judge a programme by visible engagement in its first quarter and you’re measuring the wrong group. The results, when they arrive, show up off-platform: warmer mandates, prospects quoting posts, clients raising topics first.
Don’t measure: profile-view milestones
Partly because they’re vanity numbers, and partly because of a practical truth best known early: LinkedIn does not give third parties analytics access to personal profiles. Each partner can see their own analytics tab, and nobody else can pull that data centrally. No agency, and no marketing manager, can send the firm a monthly report of every partner’s profile engagement, and anyone promising one is promising something LinkedIn doesn’t allow. The workable approach is to show partners what to look for in their own analytics tab, and let the firm-page numbers, which you can access, carry the central reporting.
The 90-day review
When the review lands, the shape of the meeting is straightforward: here are the leading indicators, here’s what they usually predict, here are the conversation moments we’ve logged, and here’s the proposal for the next 90 days. You’re not proving ROI at day 90, and you shouldn’t pretend to. You’re proving the machine works, the rhythm holds, and the early signals point the right way. The programmes that survive to month twelve are almost always the ones that were measured correctly at month three.