Zoho CRM integrations: what's a day and what's a project
"Integrate it with our system" covers everything from ticking a box to three weeks of development. The word hides the difference, and quotes vary wildly as a result.
You can usually tell which one you are asking for before anyone quotes you, and it is worth knowing, because it tells you whether a number you have been given is plausible.
Four tiers, roughly
The jump between tier three and tier four is where budgets break. Once there is no ready-made connector, you are commissioning software, and software needs specification, testing, error handling and someone to maintain it when the other vendor changes their API. A partner who quotes a custom integration in the same breath as a native one has not understood what you asked for.
The questions that matter more than the technology
Most integration problems I get called in to fix are not technical. They are these five questions never having been answered.
Which system owns each field? If a customer's address can be edited in both CRM and your accounting system, one of them has to win. Decide which, per field, and write it down.
One way or both? Two-way sync is more than twice the work of one-way, because you have to handle conflicts. Many integrations people describe as two-way turn out on inspection to be one-way with an exception, which is far cheaper.
What happens on a conflict? Last write wins, source system wins, or flag it for a human. All three are valid. Not choosing is what produces the mysterious data corruption six months later.
How often? Real-time is more expensive and more fragile than hourly. Ask honestly whether anyone would notice a thirty-minute delay. Usually nobody would.
Who gets told when it breaks? Because it will.
Middleware or direct
Zapier and its relatives are genuinely useful and I recommend them more often than clients expect from a consultant. For low-volume, simple flows they are cheaper to build, faster to change, and your own team can adjust them without calling anyone.
They are the wrong choice at volume, where the per-task pricing stops making sense, and for anything with complicated mapping or where a silent failure would cost real money. The decision is less about sophistication and more about blast radius: how bad is it if this stops working for a day and nobody notices.
What actually breaks, later
Credentials expire. Tokens have lifetimes and someone has to renew them, and that someone is often an employee who has left.
Fields get renamed. Somebody tidies up a field label in one system and the mapping silently stops matching.
Rate limits get hit. Everything works fine in testing with twenty records and falls over at month end with four thousand.
And vendors change APIs, usually with notice that goes to an address nobody reads.
None of this is avoidable. What is avoidable is finding out late, which is why I would rather spend half a day on failure alerting than add another feature to the sync.
Build it after go-live, not before
The same argument as automation generally. Integrations built before anyone has used the CRM encode assumptions about a process that has not been road-tested. Get core CRM live, let people work in it for a month, then build the connections, by which point you will know which data actually needs to move and which was only requested in theory.
The exception is anything your business genuinely cannot operate without on day one, which is a shorter list than it first appears.
Questions I get asked
Our system has no API. Options?
Scheduled file exchange, usually. Less elegant, considerably cheaper, and perfectly adequate when a daily sync is enough. Screen-scraping exists as a last resort and I would avoid it: it breaks every time the other system changes its interface.
Who maintains it afterwards?
Settle this before it is built. A custom integration is software that someone has to own. If your partner built it and then disappears, you have a dependency with no support behind it. Ask about documentation and handover as part of the original scope, not as an afterthought.
Can we do it ourselves?
Native Zoho connections, yes, confidently. Middleware flows, yes, with some patience. Custom API work, only if you have a developer who will still be there in a year. The build is the small part; the maintenance is what you are really committing to.
Disclosure: This site is written and operated by Flowpeak Consulting, which sells Zoho implementation work. Weigh the advice accordingly.
Keep reading
What it costs
Why the third integration is the expensive one, and how quotes should be broken down.
Read the guide →Implementing Zoho Books
The CRM to Books connection, and which system should own the customer record.
Read the guide →