Every business has a process it would rather not look at too closely. At CJO Accountancy it was the one that starts when a supplier invoice arrives and ends when the supplier is paid. Nobody was complaining about it. It worked. It had also, over a few years, grown to seven steps, and every one of them was done by a person.
We followed one invoice from the inbox to the bank. Not the process as the manual described it; the process as it actually ran, with the waiting, the chasing and the retyping left in.
The work as it was
- The invoice is reviewed and approved.
- It is logged on a tracking spreadsheet.
- It is re-keyed into the finance system.
- It is printed and passed across the office for a second approval.
- If that approval has not come back by Friday, someone goes and asks for it.
- The approval is typed back into the ledger.
- Payment is confirmed.
Seven steps. Two of them are the job: someone decides the invoice is right, and someone confirms the money has gone. The five in between had each been added for a reason. None of the reasons still applied.
Six decisions, in order
We put every finding through the same six decisions, in the same order, every time. Stop, simplify, buy, automate, connect, build. The order matters. Build is last because it is the most expensive answer and the hardest to undo, and because most of the time an earlier decision has already solved the problem.
Stop. The work serves no useful purpose. Remove it. The printed second approval was a check on a decision that had already been made on screen in step one. It went.
Simplify. The process is more complicated than the job requires. Redesign it. The Friday chase was a symptom of approvals with no date and no visible queue. Give the approval a date where the approver can see it and there is nothing to chase.
Buy. Someone has already solved this properly. Use their product. Nothing to buy here; the practice already paid for a finance system that could do what the process needed.
Automate. Repetitive human work that technology can absorb. Re-keying invoices into the finance system is the definition of it. The system reads them instead.
Connect. The problem exists because systems don’t talk to each other. The tracking spreadsheet and the approval typed back into the ledger were both the same fact, held in two places, because the inbox and the finance system had never been joined. Join them and the spreadsheet has no job.
Build. The right solution doesn’t exist. Create it. Nothing to build here.
The work as it is
Review and approve. Confirm payment. The five steps between them are gone, and the two that remain are the two that always mattered. No software was written. A system the practice already owned was set up to do its job, and a step with no purpose was stopped.
That is the usual shape of it. The interesting question is rarely what to build. It is why the work is still happening at all.
If you want the six decisions in full, with the questions we ask at each one, they are in the field guide, From friction to flow. If you want us to follow one of your processes, start with one process.