Order matching went from hours to instant
A meat supplier outgrew the order system we built for them years ago. We replaced it with one that matches every order to the right stock instantly. Monthly sales have about doubled since. Here is how, including what went wrong.
The problem, in their words
Every order has to be matched to a specific carcass in the warehouse. Someone checks the delivery date, finds stock that is not yet assigned, makes sure it will not expire before the order is cooked, and checks nobody else has already taken it. Done by hand with spreadsheets, that took minutes per order, sometimes hours, and some orders took days to confirm. The early system we built for them on WordPress cut that to about 20 to 30 minutes, up to an hour on a bad day. Better, but it still had no reliable way to use the oldest stock first, so people had to double-check everything.
What we saw
The team was not the problem. The system was, and it was ours. We had built that first version on WordPress, which we have done well for clients since 2015. But WordPress stores custom data in a way that gets slow when you search it constantly. This business tracks every piece of stock as its own record, not a number on a shelf count: 20,000 pieces means 20,000 records to search for each order. That is the right design for meat, and it was too heavy for WordPress. There was a second problem behind the first. The owners could not push marketing, because the warehouse, deliveries and kitchen were already stretched. More orders would only have meant more delays.
How we solved it
We rebuilt the system from scratch instead of patching it, with a simple database built for fast searching. When an order comes in, the system picks the right piece of stock instantly, oldest first, checked against the cooking date. We also split the one shared dashboard into separate views, so each role sees only its own job: the person taking orders, the warehouse, deliveries. To keep orders flowing during the switch, we did not change over in one night. The team started using it in June, old and new ran side by side, and it was fully live by September.
What went wrong
The first thing that went wrong was our own first version. It worked well for a while, until the business outgrew it. To fit the budget on the rebuild, we also agreed to skip most of the design stage and work from rough sketches. It saved money at the start and cost time later: many change requests after launch were screen details a proper design stage would have settled, like a button that existed on desktop but not on mobile, and we spent time agreeing which changes were new work and which were not. And the work grew past the original scope, which was only to fix the stock order and speed things up. We gave the extra hours for free. This client has referred other businesses to us, and we wanted to show the new way of building was worth it.
The effect
Matching an order to stock is now instant. The team still takes up to 10 minutes to double-check an order, by choice, not because the system needs it. With that time back, the business could finally take on more orders and market harder. These are their paid sales from the system, with cancelled and unpaid orders removed. We are sharing changes, not amounts.
In numbers
- Monthly sales roughly doubled from May, before the switch, to October.
- October sales were about 70% higher than October last year.
- Halfway through November, they had already reached about 90% of last November's full month.
- Sales for 2026 so far are nearly three times all of 2025.
Two honest caveats. The new system is not the only reason sales grew, and the owners did a lot of work too. And their records in our system only start in early 2025, when some sales may not have been entered in the old system at all. So the year-on-year number probably flatters us. The month-to-month jump around the switch is the one we trust most.
Where it stands now
The system is live and runs the business every day, and orders have not stopped since the switch. It keeps improving: as the team finds new ways to work faster, we discuss them together and build the changes as we go, along with the screen design work we skipped at the start. What is next: because they sell online, the BIR's e-invoicing rule applies to them, with a December 31, 2026 deadline. It came up while we were pricing another feature, and their accountant confirmed it. The paperwork is filed and we expect to make the deadline, though when the BIR approves it is not in our hands. Nobody asked us to look for it.
What we do differently now
Two things. Pick the tool for how the data behaves, not for what we already know how to build. A marketing site that mostly needs pages and a blog can launch quickly on a ready-made platform, because the editing tools already exist. A system people search and update all day belongs in a custom build. The two can sit side by side, and here they do: the main website stayed where it was, and the system runs on its own. And never skip design to save money on a system people use all day. We now include a short design stage in every custom system, even the small ones. The rest we would do again: rebuild instead of patch, switch over without stopping orders, and keep looking out for how their business is doing. Their sales show it was worth it.
Related
- Custom systems: from ₱200,000 ($3,200), scoped in writing
- Project rescue: when a system is stuck or slow
- Open proposal: one system for quotes, production and payroll
- See all case studies
Sources: KPMG: BIR RR 26-2025 extends e-invoicing deadline to Dec 31, 2026 · Grant Thornton PH: who is covered
Send us your brief and get a price.
A short brief, then a written proposal within 1 working day. We talk once the work needs it.
