Cost Per Registration: How to Defend Paid Media Spend to Finance
Cost per registration (CPR) is marketing spend divided by the registrations that spend produced. Divide all spend by all registrations and you get a blended average that no channel produced. The figure that changes a decision is cost per registration for the channel you are thinking of cutting, on the registrations that channel produced, against what a registration is worth to you.
What is cost per registration, and which version should you report?
There are four versions, and different conversations need different ones.
Paid cost per registration = Paid media spend ÷ Registrations attributed to paid media
Channel cost per registration = Spend in channel ÷ Registrations attributed to that channel
Marginal cost per registration (channel) = (Spend this period − Spend last period) ÷ (Registrations this period − Registrations last period)
Blended is the CFO's number, because it answers what acquisition costs in total. Channel is the media buyer's, because it names something you can turn up or down. Marginal is the one almost nobody calculates and the one that decides whether more money in a working channel keeps working, because it measures the last dollar where the others measure the average.
Two definition choices come first. Decide whether spend includes agency fees, creative, and staff time or only media cost. Both are defensible, and mixing them across periods is not. Then decide the denominator: all registrations, or only paid ones net of comps and staff. Write both down and keep them.
Why does the blended number hide a failing channel?
Because the average is subsidized. House channels carry almost no marginal cost and produce registrations that would largely have arrived anyway: the member email list, the past-attendee file, organic search, the speakers' own audiences. Pooled into one average, they can absorb a badly performing paid channel entirely.
Take 1,000 registrations, 600 of them from house channels at close to zero cost, and $40,000 of spend, all of it paid media. Blended cost per registration is $40,000 ÷ 1,000 = $40. Paid cost per registration is $40,000 ÷ 400 = $100. That $40 is consistent with two paid channels both at $100. It is equally consistent with one at $55 on 320 registrations and another at $280 on 80. The blended figure is identical in both cases, and only the second needs action this week.
It is also why year-over-year movement in the blended number is hard to interpret. If your house list grew, blended cost falls without any paid channel improving. Report blended, then paid, then the channel table.
Which channels does last-click credit flatter?
Three, all of them close to the registration form.
Retargeting. It only reaches people who already visited your registration page, so something else produced the visit. Under last click it collects credit for a decision made elsewhere. It is often worth running and rarely worth what a last-click report says.
Brand search. Paid search on your own event name captures people already looking for you, some of whom would have arrived through the organic result at no cost. The honest test is a controlled pause, watching total registrations.
Email to your own list. Email is usually the last touch because it is the channel with a link in it at the moment of decision. Crediting it fully makes the list look like an acquisition engine when its job is converting demand created elsewhere.
None of the three needs punishing. The correction is to stop ranking channels on last click, where the touch nearest the outcome takes the credit because it is the only one recorded. Whatever you use in its place is a model, so state which attribution model you used and keep it fixed for the season.
What should a registration be allowed to cost?
Derive it from contribution. A published cost-per-registration figure does not carry your ticket price, show rate, or cost base.
Acquisition ceiling per registration = Net revenue per registration × Target acquisition share
Ancillary uplift (add only with evidence) = Average ancillary spend per attendee × Expected show rate
Target acquisition share is a policy choice agreed with finance: the share of net revenue you will spend to fill a seat. Write the ceiling into the event marketing budget before the campaign opens.
Two refinements matter. The show rate belongs in the formula because a registration that does not turn up still cost the acquisition spend. Use your own attendance rate from the last comparable event. And if a registration is worth more than the badge because of renewal, repeat attendance, or pipeline, that belongs in the ceiling only with evidence, such as your own count of who came back.
How do you decide to cut or double a channel three weeks out?
With a rule written before the campaign starts, because a rule invented under pressure is a preference with arithmetic attached.
The Maritz Registration Insights Report analyzed more than 360,000 registration records across 30 trade shows and found that in 2023, 45% of registrants signed up in the final 4 weeks before the event, 29% in the final 2 weeks, and 9% on site (PCMA Convene). Those are trade show registration records, not a survey and not association conferences, so check the shape against your own curve. Three weeks out falls between the four-week and two-week marks, and on those shows a sizable share of registrations was still to come. A reallocation at that point can still change the room. The same analysis at the wrap meeting can only inform next year's plan.
| Decision | Condition, for each paid channel | What you do |
|---|---|---|
| Cut or pause | CPR above 1.5 × ceiling, spend at or above your stated minimum, and your minimum registrations or 2 weeks observed | Pause the channel. Move the freed budget to the lowest marginal CPR, not the lowest average |
| Hold | CPR between 0.8 × and 1.5 × ceiling | Change creative or audience, not budget |
| Increase | CPR below 0.8 × ceiling, and marginal CPR over two weeks still below the ceiling | Move budget in steps of at most 30% a week |
| Freeze | Final 5 days before the event | No changes. Delivery lag will misattribute the result |
Three details make this work. Set the minimum spend and observation window in advance, or you will cut a channel on eleven registrations of noise. Move money in steps, because a channel's cost per registration usually rises as you push volume through it, which the marginal figure detects and the average hides. And keep the rule for the whole season, including the weeks when it points at someone's favorite channel.
What breaks cost per registration measurement?
Five things, all fixable in advance and none fixable afterward.
Tracking gaps. A path that loses campaign parameters at an interstitial page, a third-party form, a redirect, or a switch of domain at payment sends registrations into "direct" and makes every paid channel look worse than it is. Test the full path with live parameters before launch.
Group and bulk registrations. One transaction covering twelve delegates, credited to whichever channel the booker touched, hands that channel twelve registrations for one decision. Report group bookings on their own line.
Comps and staff registrations. Every comped badge in the denominator lowers cost per registration and buys nothing. Exclude them and count them separately.
Offline and unattributable sources. Word of mouth, a chapter announcement, a speaker's promotion. Keep an explicit "unattributed" line. Spreading it across channels manufactures precision that does not exist.
Attribution windows that outlive the event. A 90-day click window on an event that closes registration in six weeks credits clicks that could not have produced a registration. Set the window to the campaign and state it.
Example: four hypothetical channels
Take a B2B conference with a $600 average paid registration price, a 90% expected show rate, and a $310,000 marketing budget. All figures are hypothetical. Payment and platform fees are $28 per registration and variable cost is $130 per attendee, so net revenue per registration is $600 − $28 − ($130 × 0.90) = $455. Finance has agreed a 20% acquisition share, which gives a ceiling of $455 × 0.20 = $91. The cut line is 1.5 × $91 = $136.50 and the increase line is 0.8 × $91 = $72.80. Three weeks out, with $210,000 spent, the picture looks like this.
| Channel | Spend | Registrations | CPR | Multiple of $91 ceiling | Rule |
|---|---|---|---|---|---|
| Paid search, non-brand | $84,000 | 720 | $117 | 1.3× | Hold |
| Paid search, brand | $19,000 | 610 | $31 | 0.3× | Increase, after a pause test |
| Paid social | $96,000 | 430 | $223 | 2.5× | Cut |
| Paid subtotal | $199,000 | 1,760 | $113 | 1.2× | |
| Email (house) | $11,000 | 1,240 | $9 | 0.1× | Receives cut budget, in steps |
| Unattributed | $0 | 240 | none | Reported, not allocated | |
| All registrations (blended) | $210,000 | 3,240 | $65 | 0.7× |
Marginal cost compares the latest two weeks with the two before them.
| Channel | Added spend | Added registrations | Marginal CPR |
|---|---|---|---|
| Paid search, non-brand | $6,900 | 50 | $138 |
| Paid search, brand | $1,700 | 50 | $34 |
| Paid social | $8,000 | 30 | $267 |
| Email (house) | $1,100 | 100 | $11 |
Read the blended line and nothing is wrong: $65 against a $91 ceiling. The paid subtotal is already over it at $113. Read the channel lines and two decisions appear.
Paid social is at 2.5 times the ceiling, and its marginal cost is higher still at $267, so the rule cuts it. Non-brand search sits in the hold band at $117 with a marginal cost of $138, so the response is creative and audience work.
The interesting line is brand search at $31. It looks like the best paid channel in the table, and it is the one most likely to be buying registrations the organic result would have produced anyway. A controlled pause for one week, watching total registrations, is the only way to know.
The rule sends cut budget to the lowest marginal cost, which is email at $11. Suppose $30,000 of the paid social budget is unspent. At $11 it buys $30,000 ÷ $11 = 2,727 registrations on paper, more than twice the 1,240 the list has produced all cycle. House channels saturate: you cannot mail the same list four more times and hold that cost. That is why the rule moves money in steps.
What to do this quarter
- Write the three definitions (spend inclusions, denominator, attribution window) and apply them to last year's data for a comparable baseline.
- Calculate the acquisition ceiling from net revenue per registration, and get finance to agree the acquisition share in writing.
- Build the channel table with marginal cost per registration next to the average.
- Test the full registration path with live campaign parameters and fix any step that loses them.
- Agree the decision rule, with minimum spend and observation window, before the next campaign launches.
- Book a weekly 20-minute review from six weeks out and hold to the rule all season.
Common questions
Is there an industry benchmark for cost per registration?
Not one worth using. Cost per registration depends on ticket price, audience size, market maturity, and how much demand comes from a house list, and those vary more between organizations than between industries. Build the ceiling from your own contribution math.
What if most of our registrations are unattributed?
Fix capture before you change budget. A high unattributed share usually means a tracking break or a path that drops parameters. Report the line weekly and watch whether it moves when you fix the path.
Our registrations are behind pace. Should we spend more?
Not before two checks. First, confirm you are behind at all. Kyle Jordan, director of meetings at INFORMS, told Skift Meetings: "Our old registration pacing models are not as reliable as they used to be." Second, find which part of the funnel is failing. If traffic is normal and conversion has fallen, more spend buys more of the same failure. The diagnostic order is on the page about low ticket sales.
Where EventIQ fits
EventIQ replaces nothing. It connects on top of the platforms you already run: event platforms (Cvent, Zoom, Swapcard), CRM (Salesforce, HubSpot, GoHighLevel), and marketing (Google Ads, Meta Ads, LinkedIn Ads, Mailchimp, Google Analytics). Platforms with an API outside that list are connected on request.
Marketing spend sits by event and by channel, with the author of every change. Your team enters it or imports it from a CSV. Registrations from Cvent carry their date, ticket type, and price where the platform provides them. EventIQ does not calculate cost per registration. You take the channel split of registrations from your own tracking and divide the entered spend by those registrations yourself.
Three limits change how you read it. Connecting Google Ads, Meta Ads, or LinkedIn Ads brings in campaign names and status only, so no spend arrives from the ad platforms. EventIQ does not run multi-touch attribution: an event is linked to a Salesforce campaign once a person confirms the link, and the model picker in the product does not change the calculation yet. And EventIQ does not recommend where to move budget, so the reallocation decision stays with your team.
Book a demo to see spend by event and channel next to registrations by date on a sample event, in a 20-minute demo.