Segments & tech

Custom platform or SaaS subscription: a decision framework

Five factors, one crossover calculation, and an honest map of which side your business is on. Bring a calculator, not a preference.

A 12-location chain paying $250 per location per month for its ordering stack spends $36,000 a year, or $180,000 over five years, to rent software it will never own and cannot steer. Meanwhile a single bistro that commissions a custom platform burns a launch budget on problems it doesn't have yet. Both are the same mistake: deciding by gut instead of by arithmetic. We build custom platforms for a living, and we're going to walk you through the arithmetic anyway, including the part where SaaS wins.

Factor 1: order volume is the number everything divides by

Custom costs are mostly fixed. SaaS costs scale with locations, and often with orders on top. So the number that actually decides is cost per order. At 800 direct orders a month, a $150/mo subscription runs $0.19 an order, which nothing custom can touch. Take the same tier structure to chain scale and it flips: 12 locations at $250 across 12 months, set against 30,000 orders a month, is still cheap per order, but now the fixed-cost build is cheap per order too, and it carries things the subscription never will. Rough line: below about 3,000 direct orders a month across the business, SaaS wins on price almost regardless of everything else.

Factor 2: locations, where per-location fees compound and builds don't

SaaS pricing is usually per location, so every new branch re-buys the same software. A custom platform charges once for the capability, and branch #13 costs whatever branch #3 cost, which is roughly nothing on the software side. The five-year picture for a 10-location chain looks like this, a typical $250/location/month subscription set against a representative custom build of $60,000 plus $1,500/month for platform and support:

SaaS, year 1 (10 locations)
$30,000
Custom, year 1 (build + platform)
$78,000
SaaS, 5-year total
$150,000
Custom, 5-year total
$132,000
Illustrative math: 10 locations, $250/location/mo SaaS vs. $60,000 build + $1,500/mo. Crossover lands in year 3–4: sooner with more locations, and never at all with two.

Read that caption again. At two locations the crossover never comes; at twenty it comes in year two. Location count is the steepest variable in the whole decision. And run the comparison on real quotes rather than list prices. SaaS tiers grow add-ons at chain scale (API access, priority support, extra terminals), and custom quotes hide their own tail in support and hosting. Compare five-year totals with everything included, or you're just comparing brochures.

Want this spreadsheet with your numbers in it?

Locations, volumes, current fees — we'll compute your actual crossover on a call.

Book a demo

Factor 3: data ownership, the asset that never shows up in a fee comparison

On a lot of SaaS tools you're really licensing access to your own customer base: export limits, locked segments, marketing you can only run inside their toolkit. With a custom platform the database is yours. Every phone number, order history, and address feeds your own campaigns and cashback programs, and it all follows you if you ever walk away. Price it like an asset. A base of 20,000 reachable customers generating even $5,000/month in campaign-driven orders is worth more over five years than the entire software bill above. If your growth plan leans on retention marketing, weight this one heavily.

Factor 4: roadmap control, or who fixes your bottleneck and when

On SaaS, your feature request joins a queue behind ten thousand other tenants, and the roadmap serves the median customer. You are not the median. On a custom platform, you are the roadmap. Franchise payout logic, a promo mechanic your competitors can't copy, an integration with the ERP your chain already runs: these get built because you asked. The honest counterweight is that you also pay for everything you ask for, while a SaaS product's median features arrive free with the subscription. If your workflows are standard, don't pay custom prices for standard features.

Factor 5: operational depth, or what "platform" has to mean for you

Most ordering SaaS stops where the kitchen starts. Run your own couriers and you need dispatch, zones, and live tracking (the logistics layer), plus kitchen queues and cross-location reporting, and more and more the automation sitting on top: our AI assistants alone save operators $3,000+/month on dispatch and back office and bring in $5,000+/month from campaigns. If your delivery is fully outsourced and your reporting fits inside a dashboard, this factor scores zero for custom. Don't buy depth you won't use.

Score it, then sanity-check against the two-year picture

Run the framework once: count monthly direct orders (under 3,000 points to SaaS), count locations (under 3 points to SaaS, over 8 and custom starts winning on fees alone), decide whether customer data is a marketing asset or a compliance chore for you, list the workflows no template supports, and check whether you run your own delivery. Three or more answers pointing the same way is your answer. And decide on the two-year projection, not today's snapshot, because replatforming mid-growth costs more than either option. One last correction to the old assumption: "custom" no longer means a year of development. On Dots a custom platform (apps, site, logistics, AI) goes live in about two weeks, because the platform already exists and gets shaped to you instead of written from an empty file. The timeline argument in build-vs-buy is dead; the arithmetic is all that's left.

Custom platform economics, without the custom timeline

Your brand, your data, your logistics — live in about two weeks. We'll run your crossover math first.

Book a demo

Related reading

👋 Let’s get acquainted!

Tell us about your restaurant and we’ll show you the platform on your own numbers.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.