Skip to content

Build1 publisher2 min readPublished

Running costs overtake a $1,500 automation build in seven and a half months

A shop with more than 80 automation builds published the bands it quotes in. Its monthly running cost is one flat range across all of them, so the second-year problem it warns about only shows at the cheap end.

The Engineer · Build desk

Illustration accompanying Running costs overtake a $1,500 automation build in seven and a half months

What happened

  • The automation shop 2pizza.team published the four bands it quotes in, from $500 to $1,500 for a two-app flow up to $8,000 to $25,000 for a system that replaces a manual back office.
  • Running costs are put at $50 to $200 a month for most small businesses, which the post describes as the costs that never make it into the proposal.
  • The post names four variables it says explain most of the variance between two quotes for the same brief: integration count, rules versus judgment, volume, and data quality.
  • The shop says it has seen data cleanup take longer than the automation itself, and that plenty of agencies discover that problem mid-project and come back for more money.

Compiled by The EngineerSomething wrong?How this is made

Why it matters

  • cost A buyer who signs off only the build price on a small flow is under-budgeted for its first twelve months, and the gap is larger than the invoice they approved.
  • exposure Under-declaring volume at quote time is not just a dev-hours problem: without idempotency a retry double-charges a customer, and the buyer pays that in refunds.
  • constraint Two quotes for one brief can differ honestly on four scope variables, so the lower number cannot be read as the better deal until the buyer asks which variables each side priced.
  • contradiction The post's prose puts the spread at $500 to $50,000 while its bands stop at $25,000, leaving the largest projects it claims to do without a published price.

Annualise the running band and it comes to $600 to $2,400 a year [1]. The cheapest build band, a two-app flow, is $500 to $1,500 [1]. At $200 a month, twelve months of operation costs $2,400, which is $900 more than the dearest flow the shop quotes in that band [2]. Running cost equals build cost at seven and a half months [3].

Against the top band the same figure looks like nothing: $2,400 is 9.6% of the $25,000 ceiling for a system that replaces a manual back office [4]. The same $50-200 range sits under build prices that span a factor of fifty [7].

The author writes that they run 2pizza.team, have shipped automation systems for more than 80 clients, and that the bands are what the shop charges and what it sees competitors charge [7]. "The number that surprises people is not the build, it is the second year," they wrote [6]. On the published numbers that holds at the bottom band and not at the top.

The band descriptions point at what is missing. A $500-1,500 flow is Make or Zapier configuration, two or three steps between standard apps, and the post says it "will not survive a change in either app without someone looking at it", with no error handling beyond the platform default [15]. The published figures cover the build and the monthly run, and no rate is given for that rework [8]. The post's sorting test is the useful part: "if the whole thing breaks and nobody notices for a week, you are buying configuration" [9].

The bands also overlap by construction. A judgment layer, meaning prompt work, a test set, a fallback for when the model is unsure and a way for a human to correct it, costs roughly 30 to 50% on top of the equivalent rules-only build [12]. Applied to the $1,000-3,000 multi-step band, that gives $1,300 to $4,500 [5]. The $3,000-8,000 pipeline band starts inside that spread [3].

For the bands to transfer, a buyer has to match the assumptions under them. A dashboard or admin UI is an extra $1,000-3,000 on top [16]. Error handling is priced as cheap to build in and expensive to retrofit [16]. Compliance work for GDPR or the Australian Privacy Principles is listed as scoping time added before any code [16], so a regulated buyer should read every band above as excluding it.

What to watch

  • A published hourly rate for rework after a Make or Zapier platform change would test the second-year claim at the top band.
  • Whether 2pizza.team attaches a maintenance retainer line to the same bands, and what it costs against the $50-200 monthly figure.
  • Other shops publishing their own bands would show whether $500-1,500 for a two-app flow is a market price or one rate card.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories