Team Invite

A collaboration tool used alone produces very little. The value lives in editing the same document and leaving each other notes, so with one user it is not much more than a notepad. Signups, though, arrive one person at a time.

What Theo did in his first week shows the size of that gap. He wrote a planning document and wanted his team to see it, and none of them had accounts. So he exported it to PDF and posted it in chat. Sharing it inside the product was one link away, but that link was nowhere on the screen — the invite menu sat under the third tab of the settings page. For the next two months Theo's team wrote in Pagedeck and shared through chat.

The team invite exists to close that gap inside the product. Rather than leaving the work of bringing colleagues in to whoever is diligent enough to look for it, the path gets placed where work forces people to walk past it.

Three things have to line up before an invitation actually happens: a reason to invite, a place to do it, and the first screen the recipient sees.

1. The reason gets found, not manufactured. Adding copy that encourages inviting turns the screen into an advertisement. The better move is locating the moments where the reason already exists — finishing a document and wanting someone to look at it, needing an answer in a comment.
2. The place is whatever screen that moment happens on. The same feature tucked into a corner of settings goes almost unused; placed where documents get shared, it gets used daily. The first week for someone who arrived alone through self-serve signup should route them past that spot.
3. The recipient's first screen is the last gate. Put a signup form in front of someone who clicked the link and half of them leave right there. Show the document first and ask for an account when they try to edit. The inviter's name and the document title belong on that screen too, so the visitor can tell what brought them.

Permissions and limits get decided alongside. Splitting view-only invites from edit invites lowers the cost of inviting for the sender. How many people a free plan can hold is also set here, and that limit meshes with usage-based upgrade, the design that moves a user to payment at the point a limit is hit. With five free seats, the moment someone tries to add a sixth person is the moment billing comes up — which is why the seat limit is a pricing decision and a growth decision at the same time.

Results get read as three segments: the share of signups who invited at least one person, the share of invitations accepted, and the share of accepters who edited a document in their first week. If 40 of 1,000 signups invited anyone, the first segment is leaking at 4 percent, and the place to work is where the invite button sits rather than what the invitation email says.

Products that only pay off with several people in them are the home case. Document tools, design tools, scheduling — anything where the output gets shared — make invitations the main channel of growth. When one person brought in by paid advertising goes on to bring 2.6 more, fixing one invite screen carries the same weight as raising the ad budget.

It is also the second segment of bottom-up adoption. The first segment is the door someone walks through alone; this one turns that person into a team. Going from a team to an organization is a third step, and that one is handled by admin settings and contracts rather than invitations.

Revenue attaches here too. As the number of accounts inside one customer grows, per-seat billing grows with it, which is why invitation counts get read as a revenue indicator as much as a growth one. Finance teams forecasting next quarter often start from invitation trends rather than from new signups.

Some products clearly don't fit. In a tool meant for one person — a personal budget app, a notes app — the reason to invite never arises, and a large invite button becomes a screen element nobody presses. There, a viral surface, where the output itself gets seen by outsiders, does the same job instead.

"We moved the invite button from settings to the share screen and invitations tripled."

The problem was the placement, not the feature.

"Invitations are up and the accept rate is 30 percent. Has anyone looked at the recipient's first screen?"

The bottleneck is on the receiving end, not the sending end.

"Cutting free seats from five to three will cut invitations too."

Pulling billing forward at the cost of the spread itself.

Last quarter Ravi and Noa moved Pagedeck's invite placement, taking it out of the third settings tab and putting it inside the share button at the top right of a document. Invitations sent went from 40 a week to 130.

The accept rate, though, fell from 62 percent to 34. What Mia asked about was not the invitation copy but the screen people saw after clicking the link. Before the move, most recipients were colleagues who already had accounts; after it, links started going to people who did not. Those people clicked and got a signup form instead of a document. A change that raised the count created a new problem on the receiving end, so that screen became the next piece of work.

The most common belief is that a bigger, more visible invite button produces more invitations, since not pressing it reads as not seeing it. But people invite a colleague because they need that colleague right now, not because a button looks good. A button sitting where no need exists stays unpressed in any color, and moving it somewhere useful beats restyling it. If a team puts an invite banner in the middle of the home screen and sends rise 3 percent while more than 80 percent of viewers dismiss the banner, what went up is not invitations but the number of ignored places on the screen.

The other misreading is that invitations sent equal growth, which follows naturally from sends being the number on the dashboard. Sending is only the first of three segments. One user who uploads an entire address book and fires off 200 invitations can inflate the figure, and if four get accepted, the other 196 arrived as advertising. Make sends the target and six months later the chart shows sends and spam complaints rising together.