Your best channel is invisible to your dashboard

Because the purchase runs through several people over several months, and the final click is usually a branded search. Someone reads a comparison article, mentions it in a team channel, three colleagues look it up separately, one starts a trial, procurement gets involved, and eventually somebody types your brand name into Google. Last-click credits that branded search and defunds everything that created the demand. In B2B software, branded search is a measurement of past marketing rather than a channel.

Pipeline by first touch and by assisted touch, alongside self-reported attribution on the signup form. The single most useful question you can add to a demo request is “how did you hear about us?” as free text, because it captures the podcast, the community thread and the colleague recommendation no tracking will ever see. It’s imperfect — and more honest than a model that assigns everything to branded search.

Find the activation event, then engineer everyone toward it. Almost every product has one action that separates users who convert from those who churn — inviting a teammate, connecting a data source, completing a first real task. Once you know what it is, onboarding stops being a feature tour and becomes getting people to that moment as fast as possible. Most trial funnels lose people who never reached it.

Both, with different material. The person who’ll use the product cares whether it makes their week easier. The person who signs off cares about cost, risk, security review and what happens if it fails. Bottom-up products still need a page that answers the procurement questions, because the champion has to sell you internally when you’re not in the room — and they’ll use whatever you gave them.

Enough to decide whether the business works. Recurring revenue means the customer you paid to acquire only becomes profitable after several months, so a leaky retention rate makes even efficient acquisition unprofitable. Reichheld’s research at Bain found a 5% improvement in retention increased profits by 25–95% depending on industry. In subscription software the effect sits at the stronger end, because you’re compounding.

The kind that demonstrates expertise does. The kind that restates what everyone already published doesn’t — and there’s far more of the second than the first. What earns attention now is original data from your own product, opinionated writing by people who actually built something, and documentation good enough that people link to it. Volume strategies that worked in 2018 are competing against generated content at a scale no team can match.

Legal in most cases, and frequently a poor investment. Conversion rates on competitor terms are low because the searcher already has a preference, costs are high because the competitor defends their own name cheaply, and it invites retaliation on your brand terms — which costs you more than it costs them if you’re smaller. Comparison pages that rank organically for the same intent usually return more for less.

Longer than a funding round and shorter than most people fear, provided the content is genuinely differentiated. The fastest wins are the pages you can write better than anyone because you have the data — integration pages, use cases, honest comparisons. The slowest are broad category terms where established competitors have years of authority. Sequencing matters more than volume.

Scaling acquisition before the product retains anyone. Paid channels will happily convert traffic into signups that churn in month two, and the dashboard looks like growth the whole time. The discipline is holding spend flat until cohort retention stabilises — pouring budget into a leaky funnel doesn’t fix the leak, it just makes it more expensive.