How accurate is the calculator at very large lists?
Within ±10% up to 500K subscribers. Above 200K, SaaS vendors typically move you to a custom-quote conversation with a sales team — the published step pricing is a ceiling, real quotes are 10–25% lower for serious volume. Self-hosted scales linearly through millions; the SES rate is identical at 10K/mo and 10M/mo.
Are these figures specific to Amazon SES, or do other relays match?
SES is the cheapest published rate. SparkPost adds ~$0.10/1,000; Mailgun adds $0.30–$0.70/1,000; Postmark adds $0.20–$1.15/1,000 (premium for transactional inboxing). Even on Mailgun the self-hosted path runs 60–80% lower than SaaS at 50K subscribers — the relay choice changes the savings, not the direction.
Does the calculator include automation, A/B testing, advanced segmentation?
The SaaS column reflects the Standard tier which includes those features. Mailchimp Essentials and ConvertKit's lower tiers strip them out — if you're comparing against a feature-equivalent self-hosted setup, Standard is the right column. AcelleMail ships automation (/automation), A/B testing, and segmentation at the $74 one-time price; no feature paywall.
What about transactional email — password resets, order confirmations?
SaaS marketing plans usually exclude transactional. Mailchimp Transactional, Postmark, or SendGrid run $20–$80/mo on top. Self-hosted folds transactional into the same SES bill at the same rate — AcelleMail's REST API at /api sends transactional through the same drivers. The savings on transactional are typically 70–95% by themselves.
If self-hosted is so much cheaper, why does anyone stay on SaaS?
Three valid reasons. (1) Below 1K subscribers, the SaaS free tier costs $0 and self-hosted is comparable cost once engineering overhead is included. (2) Teams without an engineer who is comfortable with SSH, DNS, and a Linux server. The abstraction tax is real and SaaS is paying for someone else's on-call. (3) Specific feature dependencies — Klaviyo's Shopify integration, Mailchimp's e-commerce data — that aren't yet replicated in open-source platforms.
How do I model this for my finance team?
Six rows: SaaS subscription (current), SaaS subscription (projected 12-mo growth), SaaS overage, SaaS transactional add-on, self-hosted total cash, self-hosted engineering line. Sum the SaaS rows; sum the self-hosted rows; multiply by 36 for three-year. The migration cost is a separate one-time row, paid from the budget that funded the first quarter of the new model.
What's the cheapest viable self-hosted setup?
A $5/mo Hetzner CX22 or DigitalOcean basic droplet plus Listmonk (free) plus Amazon SES at $0.10/1,000. At 5,000 subscribers sending weekly, that totals $7/mo cash. AcelleMail Standard adds $2/mo amortised licence and gives you multi-tenant billing, plugin SDK, automation, and a documented REST API — see /pricing.
Does pricing change if I'm in the EU vs US?
SES pricing is identical across regions ($0.10/1,000 in eu-west-1 and us-east-1). VPS pricing is identical across most regions. SaaS pricing is mostly identical, but VAT is added in EU billing — adds 19–25% depending on your country. Net: self-hosted's pricing advantage is slightly larger in the EU.
Is the calculator open-source?
The math is in the page source — view source and you'll see the JavaScript. Two functions: saasMonthlyMailchimp(subs) interpolates between published Mailchimp Standard tiers, and selfHostedMonthly(subs, sendsPerSubPerMonth) = subs × cadence × $0.0001 + $15 VPS + $2.06 amortised licence. Tweak the constants for your numbers.
What if I need a guarantee on the cost — not an estimate?
SES bills monthly per actual sends — your bill will be ±5% of the calculator estimate at any given subscriber/cadence. The variability is mostly from VPS pricing changes and bounce/retry overhead (small). For a hard ceiling, set a monthly SES sending limit in your AWS console — it caps the bill at exactly what you authorize. SaaS bills are predictable but step up at tier thresholds you may not see until you cross them.