KoltrixEstimating Transactional Email Costs as You Grow
A practical method for forecasting how many transactional emails your product will send per month, and what that volume will cost as your user base grows.
On this page(9 sections)
Transactional email is one of those costs that's negligible until suddenly it isn't. A product with a few hundred users sends a trickle of password resets and receipts. The same product with fifty thousand users, a notification feature and a weekly digest can send more mail in a day than it used to send in a year.
You don't need a precise forecast, but you do need a defensible estimate. Here's a method that takes about an hour and gets you within the right order of magnitude.
Step 1: list every email your product sends
Start with an inventory. For each automated email, note what triggers it and who receives it. A typical SaaS product sends some mix of:
- Account emails: signup confirmation, welcome, password reset, email change, login alerts
- Billing emails: receipts, invoices, failed payment notices, renewal reminders
- Onboarding emails: setup nudges, activation prompts, trial ending
- Notifications: mentions, comments, assignments, status changes, alerts
- Digests: daily or weekly summaries
- Team emails: invitations, invite reminders, permission changes
If you've never made this list, our guide to mapping every lifecycle email your SaaS sends covers it.
Step 2: estimate emails per active user per month
For each type, estimate how many a typical active user receives in a month. Use real data where you have it: your sending logs or provider dashboard will tell you the current volume by template.
Here's an illustrative example for a hypothetical B2B project tool. Your numbers will differ.
| Email type | Emails per active user per month | Notes |
|---|---|---|
| Account (resets, alerts) | 0.5 | Most users rarely reset passwords |
| Billing | 0.3 | Sent to billing contacts, not every user |
| Onboarding | 0.8 | Concentrated in new users; averaged across all |
| Notifications | 12 | The big one; depends on product activity |
| Weekly digest | 4.3 | One per week for users who keep it on |
| Team invites | 0.2 | Spiky around new accounts |
| Total | about 18 |
Notifications dominate in many products. Digests are second. Everything else is a rounding error by comparison.
Step 3: watch for multipliers
A few things quietly multiply volume:
Multiple recipients. An email to five people counts as five sends with most providers. A comment notification on a busy project might go to the whole team.
Retries. If your code retries a failed API call without an idempotency key, a transient error can send the same email twice. Use idempotency keys where your provider supports them.
Per-event notifications. One email per comment, rather than a batched summary, can turn an active afternoon into dozens of emails per user. Batching notifications often reduces cost and annoyance together.
Test and staging environments. Staging systems that send real email to real addresses consume your quota and occasionally confuse customers. Route them to a test inbox.
Step 4: project growth
Multiply emails per active user by active users for the next few quarters. Use your own growth plan; here's an illustrative example at 18 emails per active user per month.
| Quarter | Active users | Monthly emails |
|---|---|---|
| Now | 800 | 14,400 |
| +1 quarter | 1,200 | 21,600 |
| +2 quarters | 1,800 | 32,400 |
| +3 quarters | 2,700 | 48,600 |
| +4 quarters | 4,000 | 72,000 |
Add 20 to 30 percent headroom for spikes: month-end billing runs, a big customer onboarding a large team, a product launch.
Also model any feature on your roadmap that sends email. A new "daily summary" feature alone can add 20 or more emails per user per month.
Step 5: map volume to pricing
Most transactional email providers price by monthly volume in tiers, with overage charges above your tier. A few things to compare:
- Where the tier boundaries fall relative to your projected volume
- Overage pricing per thousand emails once you exceed a tier
- Whether limits are hard or soft: does sending stop, slow, or just cost more?
- What else is bundled: log retention, dedicated IPs, inbound processing, support level
As an example with real numbers, Koltrix includes API and SMTP sending in its workspace plans. Starter includes 10,000 API sends a month. Pro includes 50,000 a month, with additional sends at $1 per 1,000. The limits are enforced by the API, as described in the limits documentation. Applied to the projection above, the hypothetical product has already outgrown Starter's 10,000 sends, stays within Pro's 50,000 through the third quarter, and at the four-quarter mark would pay overage on about 22,000 emails, roughly $22 on top of the plan price, before headroom. Check other providers' pricing pages and run the same exercise.
Step 6: revisit quarterly
Set a reminder to compare the forecast with actual sending each quarter. If actual volume is running well ahead of forecast, find out why. It's often a notification that fires more than intended, a retry loop or a staging system sending real mail.
Common forecasting mistakes
Using signups instead of active users. Dormant accounts receive few notifications. Forecasting from total signups overstates volume, sometimes by a lot.
Forgetting seasonality. Billing emails cluster at month-end, education products spike at term start, and retail-adjacent products get busy before holidays. Monthly averages hide those peaks.
Ignoring per-recipient counting. Teams that grow inside your product generate more notifications per event, so volume can grow faster than your user count.
Treating the forecast as a budget. It's an estimate. Its main job is to warn you early when actual sending diverges, so you can find out why.
A forecasting worksheet
- Every automated email listed with its trigger and recipients
- Emails per active user per month estimated for each type
- Multipliers checked: recipients per email, retries, per-event notifications, staging
- Active user projection for four quarters
- Headroom added for spikes
- Roadmap features that send email modeled
- Volume mapped to provider tiers and overage pricing
- Quarterly review scheduled
Key takeaways
- Estimate volume from emails per active user per month, broken down by email type.
- Notifications and digests usually dominate; batching them can cut volume sharply.
- Watch multipliers like multiple recipients, retries and staging environments.
- Map projected volume to tiers and overage prices, and revisit the forecast every quarter.
Start with Koltrix
Your domain, one inbox, and an API that sends.
A team inbox where AI sorts and drafts (nothing is sent without your click), plus the transactional API and SMTP relay your product sends with. 7 days free, no card.
