Setting NearSync up is not one job. It is seven, they belong to different people, and only some of them have to happen in order.
The instructions for each already exist in the section that owns them. What was missing was somewhere to see all of it at once, which is this.
The one rule about order
Structure before data, data before automation.
Everything painful about a setup that goes wrong traces to breaking that.
Import ten thousand contacts before deciding your hubs, and every record lands in the wrong place. Build automations before your pipeline stages are settled, and every rule references a stage that no longer exists. Add people before permissions, and everybody sees everything on their first morning, which is difficult to walk back.
None of that is fatal. All of it is a week you did not need to spend.
The seven paths
The organisation. Your company, its hubs, its branding, its numbering. First, because everything else hangs off it.
People and access. Roles, permissions, who sees whose records, departments. Before anybody logs in.
Your channels. Email, phone, WhatsApp, web chat, your domain. Independent of the rest, so it can run in parallel.
Bringing your data in. Imports, migration, custom fields. After structure, before anything else.
Finance. Chart of accounts, tax, numbering, opening balances. Its own project with its own owner.
AI. Providers, persona, knowledge. Last, deliberately.
The checklist. One page, everything, in order.
Nothing here is irreversible
Worth saying, because the tone of this page is cautious. Almost everything can be changed later; the paths simply say which changes are cheap and which are a day's work. Do not stall on a decision you can revise.
What runs in parallel
Not everything is a queue, and treating it as one is how a two-week setup becomes six.
Channels can start on day one. Connecting a phone number and verifying a WhatsApp account involve waiting on other companies, so start them early and let them run while you do everything else.
Finance is a separate project. It has a different owner, a different set of decisions, and a dependency on your accountant rather than on anything here.
AI genuinely is last. Not because it is hard, but because an assistant configured before your data and vocabulary exist has nothing to work with, and the first impression it makes is the one people keep.
What "done" looks like
Setup has no completion event, which is why so many are abandoned at eighty percent.
A useful definition: a new joiner can be productive on their first morning without asking anybody a configuration question. They can find their work, see the records they should, send from the right address, and not see what they should not.
That is a testable bar, and it is the one the checklist is written against.
Who does what
Setup fails more often from unclear ownership than from difficulty.
One person owns the whole thing. Not to do it all, but so that "is this done" has somebody to ask.
Finance is owned by finance. A chart of accounts decided by an operations lead will be redone.
Channels are owned by whoever runs the inbox, because they are the person who finds out first when routing is wrong.
How long it takes
Honest ranges, for an organisation of twenty to fifty people.
Structure and access: an afternoon, if the decisions are already made. A week, if they are being made here, which is common and not a problem.
Channels: an hour of work and up to a few days of waiting on verification.
Data: entirely dependent on what your existing data is like. A clean export is an afternoon. A decade of a shared spreadsheet is a fortnight, and most of that is deciding what to keep.
Finance: a week, and worth taking.
AI: an afternoon, once everything else exists.
The mistake nearly everybody makes
Configuring for the organisation you plan to be.
Twelve pipeline stages because you might need them. Nine roles because the structure could grow. Forty custom fields because somebody might ask.
Every one of those is carried by every person, every day, forever. Start with what you need now. Adding a stage later takes a minute; removing one that four automations reference takes an afternoon.
What this category is not
It is not a second set of instructions. Every step here links to the article that owns it, and that article is the authority. This adds sequence, dependency and the reason for the order, which is the part nobody can see when reading one article at a time.
It is not a wizard. Nothing here tracks your progress or blocks a step. It is a map, and you are expected to skip the parts that do not apply.
Before you start
Three decisions worth settling out loud, because they are expensive to revisit.
How many hubs. Locations, entities, or business lines that need their numbers kept apart. See Hubs.
What a deal is to you. The thing you count as your pipeline. It determines stages, reporting and half the automations.
Who is not allowed to see what. Easier to open up later than to close down. See Who sees whose records.
Two things to have ready
A decision-maker who is available. Most setup delays are not technical; they are a question waiting three days for an answer about how the business works.
Your existing data, exported. Even if you are not importing yet, having the file tells you what you are dealing with, and that changes the estimate for the data path more than anything else.
If you are already running
Most people reading this are not starting empty. They have a partly configured system and a specific gap.
Go straight to the path you need. Nothing here assumes you started at the top, and each path says what it depends on.
For a general tidy-up, the checklist is the fastest way to find what was never finished.
Did this answer your question?
No, ask a person