The Pre-Show Meeting Teardown: Public Event Lists to 30 Tiered Buyers

This is the working document for turning one trade show's public lists into about 30 named buyers and booked 15-minute meetings before the doors open. It is for founders selling their own product, account executives, SDR leads and partnerships teams who have an event about six weeks out. It follows one case through every cut, gives the counts that can actually be known, marks the ones nobody publishes, and ends with a booked-versus-sent yield you can compare against your next event.

The worked case and the four public sources

The case: a seller whose buyers are VPs and Directors of Supply Chain at food and beverage manufacturers, preparing for PACK EXPO International six weeks out. Everything starts from four sources that the event or its attendees publish, and none of them is an attendee list.

• Exhibitor directory. Every company with a booth. Many of these directories run on MapYourShow, which PMMI lists among its official vendors. Exhibitors are confirmed on site, but they came to sell.
• Sponsor roster. Companies paying for tiers of visibility. The same caveat applies: mostly sellers.
• Speaker agenda. Named people attached to named sessions. This is the strongest attendance evidence you will get.
• Attendance posts. People saying on LinkedIn or under the event hashtag that they will be there. These are self-declared and dated, and some of them are about last year's show.

A few events do part of the work for you. Meet Magento Florida publishes its attendee companies with "Merchants" listed separately from "Agencies" (Wasserman, Scandiweb and others), so you get a buyer-seller split before you start. Most events do not.

Which counts can be known

This teardown uses two kinds of number, and they don't carry the same weight.

• People-level counts come from Refolk's index of professional profiles: how many buyer seats exist in each country, how many sales seats share the same keyword, and how often a keyword filter files a buyer as a seller. These are measured.
• Event-level counts are the share of an exhibitor list that is sellers, the share of attendance posts that are stale, and the share of listed speakers who drop out. I could not find any organizer or researcher who publishes them. List sellers say most "event lists" online are job-title filters pulled from generic B2B databases, often including people who never attended, but that is a claim, not a measurement. Your event produces these numbers, and the ledger in the procedure section is where you record them.

The first fork: event-first or account-first

Sources disagree on which end to start from. This guide runs event-first: harvest the event, then filter to your ideal customer profile (ICP, the account size and industry you sell to). Some practitioners run it account-first: they start from an existing target account list and check each account against the event. My rule: if you already have a worked target account list, go account-first and use the event lists only to tier attendance. If you don't, go event-first, because the exhibitor and sponsor lists will turn up accounts you did not know to look for.

Can you buy a trade show attendee list?

For most events, nobody can legitimately sell you the attendee emails. Organizers keep emails away from exhibitors on purpose and sell only blind or postal access. A third party claiming to hold the emails has no legitimate way to have got them.

AAM says it plainly: "It is AAM's policy to NEVER sell or rent attendee email contact information." h+h americas is just as firm: "There are no exceptions to this and there is no authorised reseller." AAM, ATA/TMC, PMMI, h+h americas, CloserStill's TransGlobal Events and AMTA all publicly warn that attendee-list offers are scams, and the UK Association of Event Organisers (AEO) tracks them. ATA/TMC explains where the lists come from: scammers "often use web crawler or email 'scraper' software."

This was the first wrong turn in the case. An offer arrived for a "verified" PACK EXPO attendee list. It named the event, which made it look official, and that was the false positive. PMMI calls any such offer fraudulent. The check takes two minutes: open the organizer's official-vendors page and see whether the seller is listed. One more point from AAM's warning: "responding, even with a request to unsubscribe, confirms your email address is valid." Delete the offer without replying.

Step-by-step: from public event lists to booked meetings

The procedure is nine cuts, run in order. Each cut has a count going in, a count coming out, and a rule you can state in one line. The time estimates are practitioner estimates, not measurements. Harvesting takes about a day. The ICP filter, seller removal and tiering take about half a day each. Seat mapping takes one to two days, and the CRM check a few hours. That is roughly a working week for one person, spread across the first two of your six weeks.

Keep the ledger below as the first tab of the sheet. Every cut gets one row. A count you never wrote down can't be compared against your next event.

LEDGER (tab 1)
cut,rows_in,rows_out,rule_applied,owner,date_done
01_harvest,,,exhibitors + sponsors + speakers + dated posts; deduped by company,,
02_icp_accounts,,,size and industry fit only,,
03_seller_removal,,,title-based: vendor / agency / press / competitor,,
04_buyer_seats,,,1-3 named seats per account; title variants per country,,
05_tiering,,,confirmed / self-declared / inferred,,
06_crm_collision,,,open opp or existing owner flagged,,
07_jurisdiction,,,UK sole trader / UK partnership / Germany routed off cold email,,
08_sent,,,count by tier,,
09_booked,,,count by tier,,

ROWS (tab 2)
company,person,title,country,source,source_url,source_date,buyer_or_seller,tier,crm_owner,channel,sent_date,booked,showed,opportunity

How to separate buyers from sellers on an event list

Classify every person by their title field, never by a headline keyword. In Refolk's index, the keyword "supply chain" brings in roughly one sales seat for every three buyer seats, and that is before you have touched the exhibitor list.

The second wrong turn in the case was the first pass at people. It searched headlines for "supply chain" and treated every hit as a possible buyer. The 1,046 sales seats are account executives, sales directors and business development managers at employers such as C.H. Robinson and Mainfreight Americas. Their firms sell to the 2,834, and firms like these are exactly the kind that buy booths at a show like this. The 27% is a population figure. On an exhibitor list, where everyone paid to be present, the seller share should be higher, though nobody publishes it. Count it on your own event at cut 03.

The third wrong turn came from fixing the second one. I built a seller pile by keyword and then spot-checked 25 profiles from it. Two were buyers: a Director of Supply Chain and a Senior Buyer. Keyword filtering misfiled 8% of the pile, and on a short list those are the names you can least afford to lose. The fix is to classify on the title field, then always read 25 rows of whatever you remove.

Exhibitors are confirmed on site, and that is exactly why they are the wrong people to pitch.

Exhibitor companies are still useful in three ways:

• Partnerships. If your job is selling to or alongside exhibitors, the exhibitor list is your target list, and this guide's seller cut runs in reverse.
• Account discovery. An exhibiting manufacturer can still be a buyer of your product. Its buyer seat is rarely the person staffing the booth, so that person gets mapped and tiered on their own evidence.
• Competitor removal. Competitors exhibit too. Tag them at cut 03 so nobody in your team books a "meeting" with a rival's sales lead.

How to map buyer seats per account, by country

Aim for one to three named people per account who hold the buying seat, and run separate title variants for each country. Title conventions differ enough that a filter built for US titles quietly drops most UK buyers.

The fourth wrong turn came when the case added UK accounts and reused the US query. US supply chain leaders are mostly VP-titled. UK ones are mostly Director-titled. On paper the US pool is 37.8 times the UK pool, but part of that gap comes from the query. It did not include "Head of Supply Chain", a title that is likely common in the UK, so 75 is a floor and not a market size. Before trusting any country's count, add its local titles and rerun.

This is where Refolk saves the most work. You ask for the seat in plain English, per country, with the exclusions written into the sentence. You don't have to keep a separate boolean string for each market. The UK follow-up in the case was a single sentence: "Heads of Supply Chain at UK consumer goods companies, excluding logistics providers and consultancies." Putting the exclusion in the request does part of cut 03 at the same time as cut 04.

Seat rules that held up in the case:

• One economic seat per account first. The VP, Director or Head of Supply Chain. Add a second or third seat only when there is a separate buying role, such as a Senior Buyer in procurement.
• No more than three per account. Three people at one account who all get the same "see you at the show" message will compare notes.
• The seat comes before the tier. Map who should be in the meeting before checking who is attending. Otherwise you end up with a list of whoever happened to post.

Attendance confidence tiers: confirmed, self-declared, inferred

Every name gets one of three tiers, based on the strongest public evidence that the person, not just their employer, will be on site. The tier decides the order you send in and what the first line of the message is allowed to claim.

Two of these error rates have never been published: the share of listed speakers who drop out, and the share of attendance posts that are stale. Count both yourself. For stale posts, read two things on every post: the date it was published, and the year in the hashtag or event name. A post from after the previous edition ended that says "see you next year" is fine. A post with last year's hashtag that gets reshared this year is not. Searching for people who posted within a recent window helps here: in Refolk the dated post comes back with each result, so the stale check means reading a date rather than scrolling through someone's feed.

To get from the tagged sheet to about 30 ranked names, sort by tier (confirmed buyers, then self-declared, then inferred), then by account fit within each tier, and stop at about 30. Don't keep going because the volume is cheap. Generic cold email books meetings at about 1%, so a weak list sent at three times the size still books close to nothing. About 30 is a list where every message can be written by hand.

How this goes wrong: failure modes and false positives

Most failures in pre-show prospecting come from one of two mistakes. Either a signal that proves attendance is read as proof of buying intent, or a signal that looks current turns out to be old. Each mode below lists the false positive that makes it convincing and the check that catches it.

1. Buying a "verified attendee list." False positive: the list names the event, so it looks official. Check: the organizer's official-vendors page. PACK EXPO calls any such offer fraudulent.
2. Pitching exhibitors. False positive: they are confirmed on site. Check: is the title a sales seat? The 27% seller share in Refolk's index appears before any exhibitor filtering.
3. Keyword leakage. False positive: a clean-looking seller pile built from headline keywords. Check: classify by title, and read 25 removed rows. In Refolk's sample, 2 of 25 were buyers.
4. Stale posts. False positive: an enthusiastic "I'll be there" post. Check: the post date and the year in the hashtag. The stale share is not published anywhere, so count it.
5. Country title conventions. False positive: a confident-looking UK count from a US-style VP filter. Check: run VP, Director and Head of variants for each country.
6. Treating UK named addresses as corporate. False positive: a business-looking email address. Check: Companies House status. Sole traders and other partnerships count as individual subscribers under PECR.
7. No-shows. False positive: a meeting accepted three weeks ago. As one no-show explanation put it: "The wrong person was traveling. They had to send a colleague." Check: reconfirm 48 hours before and ask who is coming.
8. Claiming attendance to inferred names. False positive: their employer has a booth. Check: the copy for inferred-tier names asks whether they will be there and never says that they will.
9. Citing 3-5x or 6-9x as fact. These are the claimed lift of pre-booked meetings over walk-ups and of next-day over week-later follow-up. False positive: both appear on many vendor pages. Check: I found no primary source with a methodology for either. Label them unverified in any internal deck.

Outreach, jurisdiction and the confirmation cadence

Send from four weeks out to one week out, run every name through a jurisdiction check before the first send, and reconfirm every booked meeting 48 hours before. One published source puts the booking window at 4 to 6 weeks before the show, and another treats the whole job as the six weeks around it.

Run the jurisdiction check on the person and their employer. The event's location doesn't decide it. This is the outline of the rules, not legal advice. Check the ICO's B2B marketing guidance and the FTC's CAN-SPAM compliance guide for current detail.

For US recipients, the CAN-SPAM opt-out must keep working for 30 days after sending. Opt-outs must be honoured within 10 business days.

The first line of each message depends on the tier, and that is where most of the personalization goes. Confirmed names get their session or booth cited. Self-declared names get their post cited. Inferred names get a question.

Subject: {event} - 15 minutes on {problem in four words}?

CONFIRMED opening: I saw you're speaking on "{session title}" at {event}.
SELF-DECLARED opening: I saw your post from {date} that you'll be at {event}.
INFERRED opening: {company} is on the {event} exhibitor list - will you be there yourself?

I'll be at {event} on {days}. I work with supply chain teams at {segment} on {one-line problem}. Would 15 minutes on site be useful? I can do {slot A} or {slot B} at {meeting point}.

If it's not relevant, no reply needed.

{your name}, {your company}
{postal address}
Reply "stop" and I won't contact you again.

48-HOUR RECONFIRM:
Hi {first name} - confirming {slot} on {day} at {meeting point}. Is it still you coming, or should I expect a colleague?

Here is what to expect, and what is not known. A widely cited cold email benchmark puts the generic reply rate at 3.43%. Another signal table shows 3.4% reply and 1.1% meeting rate for messages with no trigger. Average meeting-booked rates sit around 0.7%, with top performers at 2 to 3%. Messages that reference a trigger are reported at 5 to 18% reply. Nobody has published a reply or booked rate for event-referenced outreach specifically, or shown whether attendance tier changes it. At 1.1%, 30 cold names book zero or one meeting. The idea behind this guide is that event context plus attendance tier beats that baseline. Your event is the test.

Measuring the yield and keeping the next event cheaper

Measure three rates and keep pre-booked separate from walk-up: booked/sent for each attendance tier, show rate, and meeting-to-opportunity. A single blended number hides the result you most need, which is how much better pre-booked meetings convert.

In the one published worked example, 6 of 15 target-account meetings became opportunities and 1 of 19 other conversations did. Blended, that is 7 of 34, about 21%. Reported that way it looks respectable and hides a 40% versus 5% gap, which is the whole case for doing this work before the show. Show rate deserves the same treatment. One vendor benchmark puts pre-show booked meetings at an 80 to 90% show rate, while one customer example lost 12 of 40 to no-shows. Your number tells you whether the 48-hour reconfirm is working.

Keep these between events, because they make the next list cheaper:

• The ledger. Your counts at cuts 03 and 05 are the event-level numbers nobody publishes. Two events in, you will know your own seller share and stale-post share.
• The seller-employer exclusion list. The logistics firms, agencies and competitors you removed this time will exhibit again.
• Title variants per country. Each country's VP, Director and Head of patterns, so the UK mistake happens once.
• Booked/sent by tier. If confirmed and inferred names book at the same rate, tiering is not earning its half day, and you should cut it back.