When your business tools don't talk to each other

When your business tools don't talk to each other, a person ends up carrying the data between them: reading it in one system and typing it into another. The fix isn't always a new system. Start by finding every place the same data gets typed twice, and name who does the typing. Then check whether each tool does its own job well, and keep the ones that do. Only then decide whether to connect two tools or replace one. This article gives you a way to work through each step, a fictional example, and a checklist.

Spot the symptom: the same data typed into two systems

Most companies didn't choose their tools all at once. Scheduling came from one place, invoicing from another, inventory from a third. Each was picked because it was good at its own job. Nobody was thinking about how they would share information, because at the time a person could carry it across.

That person is still carrying it. You can usually spot the gap by these signs:

• Someone has two windows open side by side and copies fields from one to the other.
• Someone exports a file from one system, cleans it up in a spreadsheet, and imports it into another.
• The same customer has a different address or phone number in two systems.
• People ask "which one is right?" about a number that should match.
• Someone does a reconciliation at the end of each week or month to catch what didn't line up.
• Work stops when one particular person is out, because they're the only one who moves the data.

Any one of these means two systems hold the same data and a person keeps them in step by hand.

Count the re-entry points and name who does the typing

Make a list. Each line is one re-entry point: one piece of data that gets typed into a second system after it already exists in the first. For each line, write down:

• What data moves: a new customer, a finished job, a price change, a shipped order.
• From which system, to which system.
• Who does the typing: a name or a role.
• When it happens: as it comes in, at the end of the day, once a week.
• What happens when that person is out.

Count the lines. You don't need hours or dollars for this step. The count and the names tell you two things. First, how many gaps you have. Second, how much of your operation depends on specific people doing careful, repetitive typing. That second answer often matters more. Re-entry by hand is where typos come from, and the person who does it well is usually someone you'd rather have doing something else.

Test each tool: does it do its own job well?

Before you decide anything, judge each tool on its own job, separate from the gap around it. A tool that doesn't share data can still be the right tool. Ask:

• Do the people who use it every day get their work done in it without workarounds?
• If you proposed removing it, would they object?
• Is the complaint about the tool, or about having to copy things out of it?
• Does it hold data the business can't afford to lose or re-create?

If the tool does its own job well, keep it. Gap problems often look like tool problems. People say "the invoicing system is a pain" when what they mean is "I have to type every job into the invoicing system." Replace that tool and the new one sits in the same gap, with a learning curve on top.

If a tool fails its own job, that's a different problem, and it deserves its own decision.

Connecting two tools or replacing one

Once you know which tools are worth keeping, you have two broad paths. Neither is right by default.

Connecting means building the bridge between two tools so data moves on its own. It tends to fit when:

• Both tools do their own job well.
• Each tool has a way to send and receive data from outside, such as an import, an export, or an API.
• You can say which system owns each piece of data. The scheduler owns the job. The invoicing tool owns the payment.

Connecting keeps what your team already knows. Nothing gets migrated, nobody relearns their daily tool, and the change happens between the systems instead of inside them.

Replacing means retiring one tool and moving its work somewhere else. It tends to fit when:

• One tool fails at its own job, not only at sharing.
• One tool has no way to exchange data at all.
• Two tools overlap so much that one of them could do both jobs.

Replacing removes a gap, but it brings its own work: moving the data, retraining the people, and checking whether the new tool opens gaps somewhere else.

One more decision applies to either path: which direction the data flows. Some gaps need one-way flow, where one system feeds the other. Some need both systems to update each other. Settle that before anyone builds anything, because it shapes everything that follows.

Hypothetical: a scheduling tool and an invoicing tool

This example is invented. The company, the people, and the tools are fictional and describe no real business.

A residential cleaning company runs on two tools. Crews and the office use a scheduling tool to book jobs and mark them complete. The bookkeeper uses an invoicing tool to bill customers and track payments.

The symptom. Every Friday, the office manager opens both tools side by side and types each completed job into the invoicing tool. Customer addresses don't always match between the two.

The re-entry points.

1. New customer: entered in the scheduler at booking, then typed into the invoicing tool. Office manager, as it comes in.
2. Completed job: marked done in the scheduler, then typed in as an invoice line. Office manager, every Friday.
3. Rate change: updated in the scheduler's job types, then updated again in the invoicing tool's price list. Office manager, when it comes up, and sometimes missed.

Three re-entry points, one person. When the office manager is out on a Friday, invoices go out late.

The test. The crews like the scheduler and use it from their phones without trouble. The bookkeeper likes the invoicing tool. Neither complaint is about the tool. Both are about the Friday typing. Both tools pass.

The decision. Connect, don't replace. The scheduler owns customers and jobs. The invoicing tool owns invoices and payments. A new customer in the scheduler creates the matching customer in invoicing. A completed job creates a draft invoice. The office manager stops typing and starts reviewing drafts, which takes a different kind of attention: checking instead of copying.

Checklist: tools that don't talk to each other

• List every place the same data gets typed into a second system.
• For each, record what moves, from where, to where, who types it, and when.
• Note what happens when that person is out.
• Judge each tool on its own job, separate from the gap.
• Keep every tool that does its own job well.
• For each piece of data, name the one system that owns it.
• Decide which direction the data needs to flow: one way or both ways.
• Check that each tool has a way to send and receive data.
• Connect where both tools pass. Consider replacing only where a tool fails its own job.

Where the bridge gets built

You can do every step on this list without writing a line of code. The last step, building the connection itself, is where outside help usually comes in. That's what an automation build-out is for: the piece that sits between systems you already have, so information stops getting typed twice. If a tool already does its job well, it stays where it is.

If two systems in your operation keep getting typed into by hand, send us a note about which two and who does the typing.