Connecting your CRM, your production system and your accounting, without a rebuild

Nobody wants an integration project. What they want is to stop entering the same order three times. Those are different problems, and only one of them needs a big budget.

The pitch for replacing everything with one platform is always the same: one system, one database, no integration. It is a good pitch and it is occasionally right. It is also how shops lose a year and a large amount of money replacing software that worked.

The alternative is less dramatic. Keep what works, connect the gaps, and make sure no human is the integration.

The three questions that decide everything

Before any connection is built, answer these. Getting them wrong is what makes integrations fragile.

• Which system owns the customer? One of them is the truth. The others receive a copy. If two systems can both edit a customer, they will disagree and you will spend your life reconciling.
• Which system owns the order? Usually the one where the order is created. Everything downstream reads it rather than re-creating it.
• Which system owns the money? Accounting, almost always. Nothing else should be inventing an invoice number.

Once those three are settled, the connections become obvious and one directional, which is what makes them stable.

What actually needs to flow

From the store or CRM to production

• The order with its line items, decoration, placement and approved artwork
• The customer and their shipping detail, as a reference rather than a copy
• The promised date, so production is working to the same date the client was given

From production back to the customer side

• Status, in words a client can read
• The ship date and the tracking number, automatically, not on request
• Anything that changed, especially a delay, pushed rather than pulled

From both into accounting

• The invoice generated from the order that shipped, with the same line items
• Payment status flowing back onto the order, so nobody chases a paid invoice
• Supplier costs against the job, which is the only way to know your real margin

Where to start when everything is broken

Not with the hardest link. Start with the one that causes the most double entry, which is almost always order to accounting, because that one is done by hand by somebody senior and it is done for every single order.

• Week one: write down the three owners. Customer, order, money.
• Week two: agree the shared order identifier and put it in every system.
• Then connect order to accounting, because it pays back immediately.
• Then status back to the client, because it removes the most inbound questions.
• Then everything else, in the order of how much rework it removes.

Notice that the first two weeks involve no software at all. Almost every integration that fails, fails because those two decisions were never made and the connection had to guess.

When replacing everything is the right answer

Two cases, honestly.

• Your current system cannot export its own data in any usable form. That is not an integration problem, that is a hostage situation, and it is worth paying to leave.
• The pieces you are connecting each cover about a third of what you need, so you would be building and maintaining five connections forever to get one coherent process.

Outside those two, connecting beats replacing, and it beats it by a wide margin on both cost and risk.