TL;DR — Quick Summary
- Five milestones deserve a clock, not twenty: application cleared, account provisioned and logged into, equipment taking a live payment, first settled transaction, and the week-one operating review. Every other onboarding step is either inside one of these or not worth measuring.
- Measure achievement rate, not average days: “we onboard in six days” hides the tail, and the tail is exactly who leaves. Report the percentage of merchants that hit each milestone inside target, plus P50 and P90. Averages will make a broken process look acceptable.
- Publish the part you control, commit to the part others control separately: promise a business-day provisioning commitment — an input you own — and state the merchant’s obligations next to it. Promising “live in five days” promises an outcome a merchant can delay indefinitely.
What This Playbook Delivers
By the end of this, you will have a written onboarding SLA: five dated milestones, day targets for each merchant type you sell, one metric that cannot be gamed, and a fourteen-day plan to stand it up. Budget two weeks of part-time ops work — no new software required.
The Unmeasured Ninety Days
Ask an ISO owner how long it takes to onboard a merchant and the answer usually arrives as a number they have never actually calculated — “about a week,” “ten days maybe.” Ask for the distribution and the conversation stops. There is no dashboard, because the dates live in an email thread, a shipping label, and whichever rep remembers to mention go-live.
This matters more than the operational annoyance suggests. Attrition is overwhelmingly concentrated at the start of a merchant’s life: in the merchant churn benchmark, merchants in months zero to three attrite at roughly eight times the rate of merchants past year four. Those months begin in onboarding. A merchant that takes four weeks to reach its first settled transaction spends a third of its highest-risk period not yet operating, paying nothing, and forming an opinion about you.
The counter-intuitive part is how rarely this is a competence problem. The people doing the work usually know exactly which merchant is stuck and why. What is missing is an agreed clock and a shared definition of done, which means nobody can say whether last month was better than the month before — and nothing compounds without that.
1. Why Onboarding Needs a Clock Instead of a Standard
Almost every ISO has an onboarding process. Very few have an onboarding commitment, and the difference is not paperwork — it is what happens the moment something runs late.
A standard says “we aim to get merchants live quickly.” When a merchant sits eleven days, there is no exception to raise, no owner to ask, and no reason for anyone to notice until the merchant calls. A clock says “first live transaction by Day 7.” On Day 8, that merchant is an exception with a name attached, and the org either acts or consciously accepts the delay. That is the entire value: a standard absorbs lateness, a clock surfaces it.
There is also a commercial reason that rarely gets priced. Slow onboarding pushes revenue out, and pushed-out revenue is invisible: a book signing fifty merchants a month at a fourteen-day cash delay versus a seven-day one carries roughly an extra month of gross profit permanently in transit. Nobody writes that number down, because nothing in the P&L is labelled “waiting.”
2. The Five Milestones That Deserve a Clock
Most onboarding process maps have fifteen to twenty steps. That is useful for doing the work and useless for managing it, because you cannot hold twenty targets in your head and nobody will report against twenty metrics for two quarters straight. Five is the number that survives.
Each one below is written the way an SLA line has to be written: what it is, whose clock it is, and what specifically proves it happened. The done-criteria matter more than the day counts — get them wrong and the metric will be accurate and meaningless.
Application Cleared
From complete application to an approved, boarded merchant record — this is the only milestone you fully control, which is why it is the one worth publishing externally.
Software Provisioned and Opened
Provisioning is not complete when the account exists. It is complete when a human at the merchant logged in. This distinction is also what makes later software attach measurable, as covered in the attach-rate benchmark — entitlement without usage is the most common way that metric gets inflated.
Equipment Taking a Successful Test Payment
Separate “shipped” from “working.” Shipping is logistics and mostly out of your hands; configuration is yours and is where most avoidable delay sits. Only one of them belongs in a commitment.
First Live Customer Transaction
This is the milestone the rest of the merchant relationship dates from. Everything before it is setup; everything after it is retention. It is also the single day count most worth putting on a wall.
Week-One Operating Review
Fifteen minutes, deliberately scheduled, that catches everything the first four milestones cannot: a menu nobody can navigate, a close-out nobody remembers how to run, a printer that jams every evening. Skipping this is the cheapest mistake an ISO makes and one of the most expensive to clean up.
3. Realistic Day Targets by Merchant Type
A single company-wide target guarantees either a number nobody hits or a number so loose it means nothing. Split it by merchant type, because the constraint differs completely: whether someone has to build a menu, import an inventory file, or schedule across five locations.
The counts below are starting targets, not benchmarks — recommended operating goals derived from where the work actually accumulates, not from survey data. Take them as a first draft, run them for a quarter, then replace the middle column with your own P50 and keep the target where it belongs: slightly ahead of what you currently achieve.
| Merchant type | First live txn | Week-1 review | What actually takes the time |
|---|---|---|---|
| Mobile / pop-up single device, no menu |
Day 3 | Day 8 | Nothing. If this takes longer than three days, the constraint is internal. |
| Counter service / QSR one location, one terminal |
Day 5 | Day 10 | Shipping, then modifier setup and staff PINs. |
| Table service floor plan, courses, tipping |
Day 7 | Day 14 | Menu build. Almost always the merchant’s side, and why Step 4 needs merchant obligations in writing. |
| Specialty retail inventory file, barcode labels |
Day 8 | Day 15 | Cleaning up the inventory export they send you. |
| Professional services / B2B invoicing, card-on-file |
Day 7 | Day 14 | Getting their first customer onto recurring billing. |
| Multi-location (2–10) menu replication, per-site training |
Day 10 (site 1) | Day 21 | Sequencing. Set a per-site cadence of every 3 business days or the rollout drifts for months. |
Day counts are business days from application cleared, not from signed deal. Counting from signature inflates every number and quietly shifts blame onto whoever is slowest to hand over paperwork.
4. How to Measure It Without Writing Yourself a Penalty
Every ISO that considers publishing an onboarding commitment eventually asks the right question: what happens when we miss? The answer is not to avoid measuring — it is to measure the right quantity and commit only to the inputs you own.
Report achievement rate, not average days
“Average time to first transaction: 6.4 days” looks fine and is nearly useless, because it is computed across a group where most merchants are fast and a meaningful minority are catastrophically slow. Those slow ones are exactly who attrites. Instead:
Achievement rate = merchants reaching the milestone inside target ÷ merchants due in the period, expressed as a percentage. Report it per milestone, per owner, and per merchant type.
Alongside it, publish P50 and P90 days. A median of 6 days with a P90 of 31 days is a fundamentally different business from a median of 8 with a P90 of 11 — and the second one is healthier, even though the headline number is worse.
Write the exclusion list before you need it
Exclusions are legitimate. Undefined exclusions are how a metric gets discredited, because they get invented after the fact. Agree these in advance and apply them automatically rather than case by case:
- Merchant-caused delay — no staff for training, construction, permits, counter not ready, no internet circuit yet
- Incomplete application — clock starts at complete application, and “complete” has a written definition
- Underwriting hold — chargeback history, additional documentation requests
- Custom work — integrations over roughly 20 hours should move to a project track with its own timeline, not quietly break your SLA
Commit to inputs, publish targets as targets
There is a real difference between a commitment that creates liability and one that creates accountability. Structuring it correctly gets you the second without the first:
| Say this | Not this | Why |
|---|---|---|
| “We provision accounts within one business day of a complete application.” | “Your merchant will be live within five days.” | Provisioning is fully yours. “Live” depends on a menu the merchant has not built yet. |
| “Once your application is complete, deployment ships within three business days.” | “Equipment arrives within three days.” | You control when it ships. Carriers control when it arrives. |
| “Merchants are contacted within one business day, and every open item has an owner and a date.” | “All issues resolved within 24 hours.” | Response time is yours; resolution often needs a third party. |
| “Target: first live transaction by Day 7 for table service. We publish our achievement rate monthly.” | “Guaranteed.” | A published target with a published score creates accountability without creating a remedy. |
The pattern: every sentence you publish externally should describe something your team does. Anything requiring the merchant to act goes next to it as their obligation, written in the same document. That is not defensive drafting — it is the honest description of a shared process, and merchants respond well to it because it tells them what is expected of them.
5. The Fourteen-Day Build Plan
Six steps, each with a window, an owner, and a definition of done. Do them in order — step three cannot work before step two is written down.
Establish the baseline
You cannot set a target without knowing where you are. Export the last 90 days of merchants and put six dates on one row each: applied, cleared, provisioned, shipped, first live transaction, first review. If some of those dates do not exist anywhere, that is the finding — and it is why nobody could answer the question at the start of this article.
Write the five definitions and the exclusion list
Take the five milestones from section 2 and put them in your own words, with your own day targets. Then write the exclusion list. Then get it signed off by whoever owns the compensation plan — because a metric nobody is paid attention to will quietly become a metric nobody looks at.
Make the dates land automatically
Manual date entry is the reason most SLAs die in month three. Each milestone timestamp should be written to the merchant record by the system that performs it: provisioning writes the login date, the device writes the first successful auth, the processor settles the first transaction. Whatever remains manual should be one dropdown, not six free-text fields.
Split targets by merchant type and say what the merchant owes
Use the table in section 3 as the draft, then adjust to your mix. Alongside each row, write the merchant’s obligations explicitly — menu file by Day 2, staff available for a 45-minute training on Day 4, internet circuit live before install. Most merchant-side delay is not refusal, it is an unstated expectation, and once it is stated it mostly disappears.
Run a daily aging report, not a monthly review
The report has one job: list every merchant past target, oldest first, with a named next action. Ten minutes each morning. Monthly reporting tells you the quarter went badly; the aging report lets you fix this week, which is the only version that changes anything about first-year retention.
Publish what you control
Put the two or three input commitments from section 4 on the website, in the sales deck, and in the merchant agreement, with merchant obligations beside them. Publishing is what converts an internal metric into a promise your own team has to keep — and it is also a differentiator, because almost nobody in this industry publishes anything. Announcing that you will report your achievement rate monthly, and then doing it, is more persuasive than any claim about being fast.
The One-Page SLA Template
Copy this into a spreadsheet, one row per merchant. Columns 1–6 are dates your systems should write automatically; the last three are the questions you ask every morning.
| M1 Cleared | M2 Login | M3 Test auth | M4 First live | M5 Review | Days M1→M4 | Exclusion | Owner | Next action |
|---|---|---|---|---|---|---|---|---|
| date | date | date | date | date | formula | none / merchant / incomplete / hold / custom | name | text |
| — | — | — | — | — | — | — | — | — |
The three daily questions: Which merchants are past target today? Whose side is the delay on? What is the specific next action and who owns it? If your aging report cannot answer all three without someone opening another screen, it will survive exactly as long as the person who built it stays in the job.
6. Three Ways This Breaks
These are the failures that actually occur, in rough order of frequency. None of them is caused by bad tooling.
Counting from the signed deal instead of the complete application means your number includes every hour a merchant took to send a voided check. The metric looks terrible, everyone decides it is unfair, and within two months it is quietly redefined. Start at complete application and define “complete” in writing.
The single most common way this metric gets inflated. A device in a box satisfies nobody but satisfies every dashboard that defines success as dispatched. Require a settled transaction generated by the merchant, and the number changes dramatically the first month — usually for the better in the long run.
An aging report listing forty late merchants with no name beside them becomes wallpaper by week three. The correct form is one name per line and one next action. If two people can own a row, nobody does — which is also why the Uber-style escalation path (named backup after N hours) matters more than the day targets themselves.
Where OrderPin Fits the Clock
Most of Step 3 — making dates land automatically — fails because the systems that create the dates do not share a merchant record. Provisioning lives in one tool, deployment in another, first transaction in the processor’s file. OrderPin is a white-label POS platform built for ISO and MSP partners, and it puts the software side of that timeline on the same record you already work from.
- Login is a real, dated event: every merchant’s first authentication into the POS is recorded, so milestone 2 is provable rather than assumed.
- Menu build is visible from your side: you can see whether a merchant has actually built anything before their Day 4 training call, which turns a vague delay into a specific nudge.
- First transaction is observable: real settled activity, separate from test authorizations, lands on the same timeline you are reporting against.
- Software and payments on one record: the software and payment activity for a merchant sit together, so the onboarding clock and attach rate come from the same export instead of two reconciliations.
Frequently Asked Questions
What is a realistic time to first transaction?
Five business days is achievable for a single-location counter-service merchant with no menu build; seven to eight is realistic where someone has to construct a menu or clean an inventory file; ten days is reasonable for the first site of a multi-location rollout. These are operating targets rather than industry averages — set yours slightly ahead of your current median, then tighten as you find where the time actually goes.
Should we publish the SLA to merchants?
Publish the inputs you fully control and nothing else — provisioning within one business day of a complete application, deployment shipped within three business days, one-business-day response with a named owner. Do not publish outcomes that depend on merchant behaviour. Put the merchant’s obligations beside your commitments in the same document, which is both honest and effective: stated expectations get met far more often than assumed ones.
How do we stop the metric being gamed?
Fix the exclusion list in writing before you start and apply it automatically, so nobody negotiates their numbers after the fact. Require objective evidence for each milestone — an actual login, an actual settled transaction — rather than a rep’s attestation. Report achievement rate and P90 alongside the median, because a team optimizing only for the median will shorten easy cases and abandon hard ones.
Does a faster onboarding actually reduce churn?
It removes one avoidable contributor. First-year attrition has several causes — wrong-fit merchants, failed activation, competitive loss, genuine business failure — and slow onboarding mainly attacks the second. It also shortens the window where a merchant pays nothing, which is the part with the clearest measurable effect. Treat speed as necessary but not sufficient: better onboarding will not rescue a merchant who never needed the product.
What if we are too small to build dashboards?
The template above is a spreadsheet and works fine for a portfolio of a few hundred merchants. The requirement is five dates per merchant and someone looking at them each morning, not software. Automation only matters once manual entry becomes the reason rows are missing — usually around fifty new merchants a month.
Where do these day targets come from?
They are operational recommendations derived from where onboarding work characteristically accumulates — shipping versus menu build versus inventory cleaning versus multi-site sequencing — not from a survey or a proprietary dataset. Use them as a starting draft for your own process, replace the middle column with your measured P50 after a quarter, and set the target slightly ahead of that.
Onboarding decides first-year attrition, and almost nobody puts a clock on it — which is why “about a week” survives as an answer. Five milestones deserve targets: application cleared, software provisioned and logged into, equipment taking a test payment, first settled transaction, and the week-one review. Measure achievement rate and P90 rather than the average, because the average hides precisely the merchants who leave. Fix the exclusion list in writing before you start it, require objective evidence for each milestone, and commit publicly only to the inputs you own while stating the merchant’s obligations beside them. Six dependent-at-risk steps get you there in fourteen days, and none of them requires new software — only five dates per merchant, named owners, and someone reading the aging report each morning. OrderPin is a white-label POS platform built for ISO and MSP partners: first login, menu build and settled transactions all land on one merchant record, so milestone two through four is evidence rather than a phone call.
Assumptions & How to Adapt
Nature of the content: this is an operating framework, not a data report. The day targets are recommended starting values derived from where onboarding work characteristically accumulates, not measured benchmarks — no survey or proprietary dataset was used, and no figure here should be quoted as an industry statistic.
Scope: US merchant acquiring, SMB book. Day counts are business days measured from a complete application, with “complete” defined by your own intake checklist. Multi-location timelines assume a per-site rollout cadence and should be re-derived if you commission sites in parallel.
Related reading in this series: the first-year attrition concentration figures come from Merchant Churn Benchmarks; the definition of “entitled versus actually using” is carried from Software Attach Rate. Both are modeled rather than surveyed, so treat them the same way: replace them with your own measurements as soon as you have them.
About OrderPin
OrderPin is a white-label POS platform built for ISO and MSP partners. We offer full data ownership, flexible pricing, and seamless API integrations to help you build a recurring revenue business under your own brand. Learn more about OrderPin’s white-label solution

