Where the hours actually go, what order to fix things in, when to build and when to buy, what it costs and what it returns — with four real builds and twelve calculators as the evidence rather than adjectives. Written by the person who did the jobs before rebuilding them.
Most trade businesses do not have a software problem. They have a rules problem: the prices are in one person's head, the order of work is in another's, the stock truth is in a spreadsheet nobody trusts, and the owner is the integration between six apps that do not talk to each other. Systemising is the work of writing those rules down somewhere that applies them the same way every time. This guide is the method, in the order it should be done, with the evidence.
It is a ladder with four rungs, and most businesses are on the first. Paper: orders on a pad, quotes on a calculator, the schedule in someone's head. Digital: the same things in spreadsheets and folders, which is better than paper and worse than people think, because a spreadsheet only knows what somebody remembered to type. Connected: the tools talk to each other, so a quote becomes a customer record becomes an order becomes a dispatch without anyone carrying it between screens. Rules embedded: the system knows the price, the stage, the next action and who owns it, so nobody has to ask. AI goes in above that, where it earns its place — and most businesses have three rungs to climb before it does.
The test at every rung is the same: can a normal day happen without the owner? If the answer is no, the rules are still in a head, whatever software has been bought.
Five places, in every business I have worked inside. Each one has a calculator that puts your own figures on it.
Before anything clever. A business on paper cannot be connected, because there is nothing to connect. The first move is always the same: the orders, the quotes, the schedule and the customer record into something digital, even if it is imperfect. Everything after depends on it.
Usually quoting. Fixing the front of the chain first is what makes the rest possible: everything downstream then works from a clean, priced, unambiguous order rather than something typed out twice. At the nursery the unlock was not software but advice — put every product line on the website with its price — and once that catalogue existed it became the rule the quoting system follows, so nobody could underprice and a supplier rise reached every quote at once. Quoting went from fifteen minutes to two or three, on the call.
Follow the same job through fulfilment, stage by stage, to the point it is delivered and closed. Five roles touch a job on the way — sales, order processing, procurement, the yard or nursery team, logistics — and each gets its own pipeline, where every stage is a task with an owner and the data that stage needs. Take the single heaviest piece of admin next, whatever it is: at the nursery it turned out to be an office person typing every courier consignment by hand at six or seven minutes each, which became nothing the moment the outside team could process the order on a phone.
Most operational software fails at the changeover, not at the feature list. A business that cannot afford a bad week will not risk one, so it stays on the folders forever. The answer is to put the new system in while the old way carries on — the same orders through both — and stop the old way only once the new one has proved itself on real work. No downtime, and nothing at risk.
Three routes, and the right one is usually a mix. Buy where a product already knows the rules: CRMs, accounting, calendars, courier booking. Connect what you already have — the option most businesses miss, and the one that rebuilt the nursery with WooCommerce, Pipedrive, the courier portals it already used and a connection layer of Zapier, Make and a little custom code. Build only the part no product knows: a fence take-off that derives posts and gravel boards from a drawing, a plant name that has to be learned, a lorry that fills against a value threshold, a scheduler that knows driver-hours law. The full decision, with the costs of each route, is in the build-or-buy guide.
| Build | What it replaced | What moved | From |
|---|---|---|---|
| Delivery route planning | Four tools glued by retyping | 2 hrs → 10 min per run; 10–18 hrs a week back | £5,000 |
| Procurement pipeline | A workbook and an inbox | 3 → 100+ growers in one press; 14–19 hrs a week back | £12,000 |
| Fencing quoting | Paper survey, calculator, hand drawing | Drawing 1–2 hrs → 0; live in 11 days | £8,000 |
| Wholesale order management | Website-and-spreadsheet quoting; orders in folders | Quoting 15 → 2–3 min; £1m → £3.2m, office unchanged at 3 | No bespoke software |
Against UK agency rates of £50,000–£200,000+ over three to nine months for comparable scope, and off-the-shelf subscriptions that leave the spreadsheet alive beside them. The return is the other side of the same number: what standing still costs, which the first calculator works out from eleven figures you already know. The engagement itself starts with the operations audit — £1,500, two days inside the working week, guaranteed to find at least ten recoverable hours a week with evidence or refunded — and continues in sprints quoted against a named outcome.
Two tests, and neither is a feature list. Throughput without headcount. The nursery went from four people to fifteen and £1m to £3.2m, but every hire went outside, onto the nursery and the lorries; the office stayed at three. The same three people carried more than three times the revenue, which is only possible if the administrative work per order collapsed. The Tuesday test. Can a normal day happen without you? If the answers to the day's questions now live in a system — a stage, an owner, the data it needs — you are consulted for decisions rather than facts, and the business has become the thing a buyer would call transferable. The valuation calculator puts a figure on that.
Every one loads with a real business's figures, shows its formula, and copies as a link. Cost of standing still · True cost of an employee · Charge-out rate · Margin leak · Owner dependency · Valuation and dependency · Quote response time · Quotes to target · Supplier reach · Route planning cost · Late payment cost · Fence materials. All twelve are at /tools/.
Count something for a week rather than guessing it: minutes per quote, questions a day that only you can answer, days customers take to pay. Put it in the matching calculator and send the number to whoever signs the cheques. If it is worth acting on, the audit finds the hours with evidence inside your working week, and the short version of how the work runs is on the byHoratio homepage.
To move the rules the business runs on — prices, stages, priorities, who does what next — out of people's heads, paper and spreadsheets and into systems that apply them the same way every time, so the team can run a normal day without the owner and the owner can run the business rather than operate it. It is not the same as buying software; software the team never uses systemises nothing.
With whatever feeds the operation, which is almost always quoting, and before that with getting off paper. A quote that is fast, priced from one book and produced where the customer is makes everything downstream cleaner, because fulfilment then works from a clean, priced, unambiguous order rather than something typed out twice. Then follow the same job downstream, one bottleneck at a time.
Only where no existing product knows your rules. The wholesale nursery in case study 004 was rebuilt with WooCommerce, Pipedrive, the courier portals it already used, and Zapier and Make to connect them — no bespoke application at all. The fencing contractor needed a custom pricing engine because no product knows what a gravel board is. Most businesses need mostly the first and a little of the second; the build-or-buy guide sets out the decision.
Days to a working first version, not months of discovery, when the work is done from inside the business against real jobs. The fencing system was live in production eleven days after the first commit. The nursery's quoting, CRM and order management rebuild went in over three to four months, with no downtime, because it ran alongside the old way until it had proved itself.
Two tests. Throughput without headcount: the nursery went from £1m to £3.2m with the office unchanged at three people, which is only possible if the admin per order collapsed. And the Tuesday test: can a normal day happen without you? If the answers to the day's questions now live in a system rather than in your head, it worked.
The operations audit is £1,500 for two days inside the working week, guaranteed to find at least ten recoverable hours a week with evidence or refunded. Builds are quoted per sprint against a named outcome; the four documented builds started from £5,000, £8,000 and £12,000, against UK agency rates of £50,000–£200,000+ for comparable scope. The calculators show what standing still costs, which is the number to compare against.
If not, the rules are still in your head — whatever software you've bought. The audit finds them and writes them down. Twenty minutes, no deck, no pitch.
HORATIO@BYHORATIO.COM
When a product knows your rules, when connecting what you have is enough, and when to build.
Delivery planning, procurement, fencing quoting and wholesale order management, with the arithmetic.
Your own figures on the hours, the hire, the margin, the owner and the value.