The case for a spreadsheet is that it is free and it does what you tell it. Both are true. The problem is the second half: it only ever does what you tell it, every single time.
Count the hours honestly
Take a normal week and count, per order:
| Task | Typical time |
|---|---|
| Re-typing customer and address details | 3 to 5 min |
| Building the price from scratch | 5 to 15 min |
| Producing and sending a proof | 5 min |
| Chasing approval by hand | 5 to 15 min |
| Answering "where is my order" | 3 to 8 min |
| Re-typing the order into an invoice | 5 min |
| Working out what to order in blanks | 5 to 10 min |
That is 30 to 60 minutes per order of pure administration, none of which decorates a garment. At 40 orders a month and a £30 loaded hourly cost, that is £600 to £1,200 a month, most of which is your own time as the owner.
Then count what falls through
Harder to measure, more expensive:
- Quotes never followed up. A conversion rate you cannot see cannot be improved.
- Approvals sitting for days because nobody was chasing. See cutting approval time.
- Jobs late because nobody saw them coming. A spreadsheet does not raise its hand.
- Prices drifting between quotes and between staff, because the arithmetic is redone each time.
- Reprints with no approval trail, so you absorb costs that were not yours. See handling a reprint request.
- Reorders quoted from scratch because last year's job is unfindable.
Where a spreadsheet genuinely wins
Be fair about this. A spreadsheet is the right tool when:
- You do a handful of large repeat jobs a month for two or three customers.
- One person handles everything and holds it all in their head.
- Your pricing is genuinely simple, one method, few variables.
- You are testing whether the business works at all.
If that is you, buying software is premature and you should keep your money.
The signals that you have passed the point
- You have re-typed the same customer's address more than once this week.
- You have chased an approval by text.
- You cannot answer "what is due Friday" without opening two files.
- Someone has asked you for a status you had to go and look up.
- Two quotes for similar jobs came out at noticeably different prices.
- You have discovered a stockout on the morning of a print.
Three or more of those and the arithmetic has already decided.
Move in an hour, not a project
The four costs a sheet cannot see
Hours are the visible cost of manual working. These four are larger and quieter.
| Hidden cost | How it shows up |
|---|---|
| Stale prices | A quote built from last year's blank cost, won and produced at a loss |
| Unrecorded labour | Every margin figure flattering by 15 to 30 percent |
| No audit trail | Approval disputes decided by whoever remembers hardest |
| Version drift | Two people quoting the same job differently in the same week |
The second one deserves emphasis because it distorts everything else. If labour is not captured, your best-looking jobs are frequently your worst, and you will keep chasing the work that is quietly costing you. A shop that starts measuring hours against jobs, even crudely, usually reorders its own priorities within a month; the job costing view is the version that requires no analysis.
Work out your own number in ten minutes
Five figures, and be honest about them:
- Orders last month. Count them.
- Admin minutes per order. Quoting, re-keying, chasing approval, ordering blanks, status answers, invoicing, chasing payment. Most shops land between 45 and 120 minutes.
- Your loaded hourly rate. Wages plus on-costs, for whoever does that work.
- Orders that went wrong last month for a reason a system would have caught: wrong version printed, missed date, forgotten reorder, unbilled job.
- Average order value.
Admin hours times rate, plus the cost of the failures, is your monthly cost of staying manual. At 40 orders and 70 minutes each, that is 47 hours a month. At £22 an hour it is over £1,000, before a single mistake is counted.
What the transition actually feels like
The honest version: the first week is slower, because you are learning where things live and importing history. The second week is level. From the third week the difference is not that any single task is faster, it is that a category of tasks stops existing: nobody re-types an address, nobody hunts for the latest proof, nobody wonders whether that job was invoiced.
The two habits that decide whether it sticks are unglamorous. Enter every job in the system, including the small ones, because a partial system is worse than either extreme. And record stage changes as they happen rather than in a batch at the end of the day, which is why the recording has to be a scan or a tap rather than a form.
Keep the spreadsheet for what it is good at
This is not an argument against spreadsheets. They are the best tool available for one-off modelling: testing a price change, sketching a capacity plan, running the arithmetic in benchmarking your own pricing. What they are bad at is being the operational record for a process with multiple people, dates, files and money attached. Use them for thinking, not for remembering.
Migrating from spreadsheets is the easiest migration there is, because a spreadsheet is already tabular. Customers, products and historical orders paste into an import template. Most shops are done in under an hour. See switching print shop software for the harder case of moving between systems, and our spreadsheet comparison for the direct version of this argument.