THE SHORT VERSION
- Import into a fresh sub-account whenever the target account is already live.
- There are always fifteen or so client-specific fields to rewire; work from a list, not from memory.
- Nothing is installed until you have run the customer journey yourself and watched it fire.
- Turn campaigns on in order, not all at once — a misfiring blast is harder to explain than a slow launch.
Most snapshot disasters are not technical. They are a sequencing problem: someone imports, flips every campaign on, and discovers three days later that the SMS is going out signed by the wrong company from an unregistered number.
Here is the order that avoids that.
Before you import
Five checks, all of which take a minute and all of which are cheaper now than later.
- Is the target sub-account fresh? If it already has live workflows, do not import on top. Create a new sub-account, import there, and migrate what you want deliberately.
- Does the plan support what the snapshot uses? Memberships, communities and some AI features are plan-gated. A snapshot that references them will import with dead links on a plan that lacks them.
- Is the sending domain set up? Not required to import, required before anything sends. Doing it now means you are not blocked at the test stage.
- Is a phone number provisioned and A2P registered? Registration is the single most common launch delay and it is not in your control. Start it first.
- Do you have the client’s real details to hand? Name, address, service area, hours, booking preferences, pricing, and the offer the funnels should sell. Chasing these mid-build is how a two-hour job becomes a two-week job.
The import
Open the share link while logged into your agency, choose the destination sub-account, confirm. It will run for a minute or two and then report what it created.
Read that report. It lists every asset. If the count does not match what the snapshot promised, stop and find out why before you build on top of it.
The fifteen fields
Every snapshot has a finite set of client-specific values, and every good snapshot lists them. Work through the list rather than hunting. In practice it is nearly always some version of this:
| # | Field | Where it bites if you miss it |
|---|---|---|
| 1 | Business name | Every email and SMS template |
| 2 | Business phone | Templates and funnel headers |
| 3 | Business address | Email footers and compliance |
| 4 | Service area | Funnel copy and qualification logic |
| 5 | Hours of operation | After-hours routing branches |
| 6 | Booking calendar owner | Every booking link in the snapshot |
| 7 | From-name and reply-to | Deliverability and reply handling |
| 8 | Sending domain | Deliverability |
| 9 | Phone number | Every SMS action |
| 10 | Pricing / offer strings | Funnel and template copy |
| 11 | Brand colours and logo | Funnels and the client portal |
| 12 | Pipeline stage names | Reporting and stage automations |
| 13 | Team members and assignment | Task routing |
| 14 | Notification recipients | Internal alerts |
| 15 | Seasonal / date triggers | Campaigns that fire at the wrong time |
Most of these live in custom values in a well-built snapshot, which means you set them once in one place. If you find yourself editing the same string in nine templates, the snapshot was not built for redeployment — which is worth knowing about the vendor.
Test like a customer, not like a builder
This is the step people skip and it is the only one that proves anything.
Do it in this order, with a real device and a real email address:
- Submit the main funnel form. Not the preview — the live URL.
- Watch the contact appear with the right fields populated and the right tags.
- Receive the SMS. Check the sender, the wording, and whether the opt-out language is present.
- Receive the email. Check the from-name, the reply-to, the footer address, and whether it landed in the inbox rather than promotions.
- Book the slot. Confirm it lands on the right calendar with the right owner and the right buffer.
- Move the pipeline stage. Confirm the stage automation fires.
- Wait for the second touch. Most cadences have a delay step; if you never wait, you never test it.
Record this. A two-minute screen recording of the journey working is the single most useful artefact you can hand a client, and it removes an entire category of “is it actually live?” conversation.
Switching campaigns on
Not all at once.
Turn on the intake and the immediate-response workflows first, and leave them running for a few days on real traffic. Then turn on the follow-up sequences. Then, last, anything that touches an existing list — reactivation, seasonal blasts, recall campaigns. Those are the ones that reach hundreds of people at once and they are the ones you want running last, when everything upstream is proven.
If you are working a large old list, respect the quiet-hours and daily-cap logic rather than removing it to go faster. Carrier filtering is unforgiving and getting a number flagged costs more than the fortnight you saved.
When to hand it over
If you are installing more than two or three snapshots a month, the maths on doing it yourself stops working. A snapshot installation turns a variable four-hour job into a fixed-price 72-hour one with a recorded test at the end, and if you are onboarding a whole agency rather than one account, white-glove onboarding is the wider version of the same trade.
Either way, the rule holds: nothing is installed until you have watched it work.