What five event programs taught us about event outbound

Over roughly a year we ran the pre-event and post-event outbound for an identity-document AI company across a set of third-party events. Five of them are worth writing about, because between them they cover the Gulf, Asia, Europe and Latin America, plus one deliberate attempt to take the motion into a new industry.

I am leaving the numbers out. The per-event counts belong to the customer and are not mine to publish. The lessons underneath them do not need the counts.

1. Channel performance is regional

At the Gulf fintech summit, LinkedIn clearly out-performed email, and the gap was not marginal. The same target list, the same persona segments, the same week, and the replies came through one channel and largely not the other.

In Latin America the sequence needed a messaging app in it to work at all, which is not a channel that appears in any of the standard B2B outbound templates. And the first Portuguese copy we produced was written directly in Portuguese rather than translated from the English, because translated outbound reads as translated outbound and gets treated accordingly.

In Europe, email carried most of it.

In practice: you cannot decide your channel mix once and export it. Treat your home-market instinct as a hypothesis in a new region, and test it in the first week of the campaign rather than discovering it in the last one.

2. Mid-tier companies out-replied the large enterprises

At the APAC banking summit, the mid-tier fintechs out-replied the large banks. Consistently, on the same sequences, at the same event.

That was not what we expected going in. The named-account list leaned heavily on the large institutions, because that is where the deal sizes are, and the effort followed the list. What came back was that the smaller and mid-sized companies were both faster to answer and more likely to say something useful when they did, including when the answer was no.

One explanation is that a mid-tier company has fewer layers between the person reading the email and the decision. Another is that the large institutions were being sold to by everyone at that event and the mid-tier ones were not.

What we changed as a result was not the target list. Large banks are still worth a seat. What changed was the expectation attached to each segment, and the willingness to adapt mid-flight when early replies from one segment came in ahead of the other.

First-week reply data in a 30-day sequence costs nothing to read, and it usually goes unread until the retro.

3. Framing mattered more than channel with enterprise buyers

At the European banking summit, the messaging that named the systems the bank already runs and promised to sit alongside them out-performed messaging built on replacement. Same event, same titles, different framing.

It is not obvious when you are writing the copy, because the sender is usually proud of the thing their product replaces and the replacement story is the one the sales deck is built on.

A risk or compliance leader at a large bank hears a replacement story as a project: an integration, a migration, a security review, an internal fight, and a year. A “this sits next to what you have” story is a conversation they can have without any of that. The change they have to picture is a much smaller one.

If you sell into regulated enterprises, that framing is worth testing as a variant before you test anything else in the copy.

4. Vertical fit does not transfer

The most instructive failure in the set was the hospitality technology trade show.

The logic going in was sound: the product solves an identity and document problem that hotels also have, the event was well attended, and the ICP filter produced a list that looked reasonable on paper. The outreach went out and mostly went nowhere.

What we got wrong was not the vertical. It was that the copy talked about the problem in the language of the vertical we knew, and hospitality buyers evaluate everything through the systems they already run, particularly their property management system. Messaging that did not connect to those systems had nothing to attach to.

The buyers also had less urgency than the financial services buyers, which is a separate and equally real finding. Urgency is part of ICP fit, and firmographics do not show it.

Two lessons came out of that event. A new vertical needs its own domain-specific messaging before the first sequence goes out, not after the first sequence underperforms. And an ICP hypothesis is worth testing at a smaller event before it is tested at an expensive one.

5. Event apps are unreliable

Across all five events, the event app was the least dependable part of the plan.

• At one event the in-app messages went largely unread.
• At another the platform was closed enough that building a usable attendee picture from it was impractical, so the picture had to come from the open web and the platforms we were entitled to instead.
• At a third, the app worked well enough to be genuinely useful for meeting requests on the floor.

The correction is not to ignore event apps. It is to find out which kind you are dealing with four weeks out rather than the week of. If in-app messaging is a real channel at that event, plan for it. If it is not, you have four weeks to build the meeting target out of email, LinkedIn and named-account overlap instead of finding out on the Monday that a third of your plan does not exist.

The expensive version is finding the app unusable late and leaving the meeting target where it was. The program then looks like a copy problem in the retro.

6. Starting late depressed conversion at every event

This is the only lesson that showed up at all five events.

Outreach that started close to the event day underperformed at every one of them. The reasons stack:

• The recipient’s attention has already moved to travel and agenda
• Their inbox is at its fullest
• The sequence has to be compressed into a few days so it reads as pushy
• There is no room left to revise the copy based on what the first replies tell you

None of those are fixable with better writing, because they all follow from one date.

The launch date is the variable worth defending. Most things in an event program can be renegotiated or fixed afterwards. Four weeks of runway cannot be recovered in week three.

What we do differently now

The six lessons above turned into four operating rules.

1. Outreach launches the same week the attendee list is built, on the first enriched batch rather than the complete one.
2. Channel mix is a per-region question with a first-week test attached.
3. New verticals get their own messaging written before the first send, sourced from someone who knows the systems that industry runs on.
4. The event app gets assessed at four weeks out, with the meeting target revised on the spot if it turns out not to be a channel.

Those rules are the basis of the Event Outreach Framework, which is published in full with the copy examples and the ownership split. If you want the wider motion around the outreach, the tiering model and the post-event triage service levels are in The Event GTM Playbook.

Frequently asked questions

Is LinkedIn or email better for pre-event outreach?

It depends on the region and it is worth testing per market rather than deciding once. In the Gulf events we ran, LinkedIn clearly out-performed email. In Latin America a messaging app became a necessary part of the sequence. In Europe email carried most of the replies. The mistake is exporting a channel assumption that worked in your home market to a region where the habits are different.

Are event apps useful for pre-event outreach?

Treat them as a bonus rather than a channel you plan around. Across these five events the in-app messaging was either largely unread or the platform was closed enough that getting a usable attendee picture out of it was impractical. Find out which kind of app you are dealing with four weeks out, not the week of, so you can size the rest of the plan accordingly.

How late is too late to start pre-event outreach?

Inside two weeks the results fell off at every event we ran. A 30-day window lets a five to seven touch sequence sit two to three days apart. Compress that into six days and it reads as urgent for the sender rather than useful for the reader, and the reply rate follows.