The only two kinds of system worth building

Two categories, no third

A revenue driver makes the company sell more, or sell the same thing to more people without more effort. Speed to lead, follow up that never stops, referrals that get asked for on time, reactivation of clients who went quiet.

A streamline system makes existing work cheaper, faster or less error prone. Invoicing that runs itself, onboarding that collects its own documents, reporting that arrives before the meeting.

Anything that fits neither is noise. It might look impressive in a demo. It will not show up in a P and L.

How to tell which one you are being sold

Ask a single question: which line of the accounts does this move, and in which direction?

Revenue drivers move the top line. Streamline systems move cost, or move errors that turn into cost. If the answer is better visibility, keep asking. Visibility is not a result, it is a step towards a decision somebody still has to make.

Better visibility is what people say when a system does not have a number attached to it yet.

Why smaller companies should start with revenue

A 25 person company saving 30 hours a month has 30 hours of somebody's time back. Useful, and it does not pay for the next project.

The same company adding four deals a month has budget, and more importantly it has proof. Internal permission is the scarce resource in the first year of changing how a business runs.

Larger operations flip this. At 300 people, three percent off delivery cost is a bigger number than most sales improvements, and the machinery to capture it already exists.

The arithmetic, side by side

One 30 person agency was choosing between two projects with almost identical build cost.

Option A automated the reporting pack, saving 26 hours a month at $73 an hour. That is $1,900 a month, a real number and a genuinely nicer month end.

Option B was follow up: sequences for proposals that went quiet, plus reactivation of clients dormant for 90 days. Modelled conservatively at two extra deals a month against an average of $7,000, that is $14,000.

Both were worth doing. The order changed the year, because option B paid for option A by week eight.

The trap in that comparison

Time savings feel certain and revenue feels speculative, so committees pick option A. Be conservative on the revenue model instead of avoiding it. Half of a well modelled number still beats all of a small one.

Where the boundary blurs, usefully

The best systems do both, and they are worth looking for. Speed to lead is the obvious case: it wins deals and it deletes the chasing.

• Automated onboarding shortens time to first invoice while removing admin.
• A CRM that fills itself gives sellers hours back and surfaces who is ready to buy again.
• Reactivation campaigns cost nothing to run once and reach people your team has no time to call.

When a candidate lands in both columns, build it first. Those are rarer than vendors claim.

What this rules out

Applied honestly, this framework kills a lot of popular projects.

• The internal portal that centralises information nobody was struggling to find.
• The executive dashboard with 40 metrics and no decision attached to any of them.
• The chatbot on a site with 200 visitors a month.
• The custom platform to replace five tools that work, built because integration felt untidy.

Each of those has been proposed to us by a client with a straight face. We say no, explain the arithmetic, and offer the version that has a number on it.

How to use this on Monday

Take your current list of ideas. Write R or S beside each one, then the monthly figure you expect and how you would measure it.

Anything you cannot label is not ready yet. The lines carrying an R and a credible number are your next quarter.

How to model a revenue system honestly

Revenue projections get dismissed because most of them deserve it. Build yours from three inputs you already hold and it survives scrutiny.

1. Current volume: leads, quotes or clients touched per month. Use last quarter, not your best month.
2. Current conversion at each step, taken from your own records rather than an industry benchmark.
3. One change, stated conservatively. Half the improvement you expect is a sensible default.

Multiply through, then take 20 percent off for the fact that reality is untidy. When the result still pays for the build inside a quarter, you have a project you can defend to anybody.

The order in practice

Most clients run three or four systems in the first year. A typical sequence starts with speed to lead, adds follow up and reactivation, moves to onboarding, then finally to reporting.

Reporting comes last on purpose. It is the most requested and the least urgent, because a report changes nothing until somebody acts on it.

One exception is worth naming. When leadership cannot see the numbers at all, a small reporting piece goes first, purely so everything after it can be measured.