Marketing Engineering for Local Businesses: The First System to Build

What should a local business automate first?

In High Output Management, Andy Grove books dinner by phone. The woman who takes the call is new. It seems nobody told her the restaurant had lost its liquor license.

So nobody brings wine. The maitre d' walks from table to table, apologizing, all night long.

One missing sentence on one phone call. Every guest in the room pays for it.

That is what marketing automation for small business is really about. Not more emails, not a fancier website. The handoff from one step to the next.

A local service business rarely has a lead problem first. It has a dropped-handoff problem: the call nobody returned, the quote nobody sent, the patient nobody rebooked.

The proof is ugly. In a secret-shopper study by the legal software company Clio, only 40 percent of law firms picked up when called. Almost half could not be reached by phone at all, even after a message.

Those callers had already found the firm, picked it and dialed. The marketing worked. The handoff died.

Marketing engineering, sized for a 20-person shop

Marketing engineering is building the systems behind your marketing, so the next step happens because a system does it. The term began in a 1998 marketing textbook as a way to make decisions with data, models and judgment.

That textbook already said simple models stay useful for small firms without big databases. Good news. You do not need a data team.

A marketing engineer also builds the systems that make those calls every week. At a software company that means pipelines and agents. In a shop with three trucks it means four or five plain rules.

Every system we build has the same five layers: signal, memory, decision, action and feedback. Here is the small version.

Signal is your call log, your calendar and your reviews. Memory is the price list, the service menu and what each customer told you. Decision is the rule, like who gets a callback and when a quote goes out.

Action is the text, the quote or the reminder. Feedback is one number you read every Monday.

Do you need a full-time hire for that? Usually not. You need one person who owns each system once it runs, and approves anything a customer will see.

A system counts when it still works next Tuesday with nobody remembering it.

Map your seven handoffs from last month's records

The Handoff Map is a one-page list of the steps a customer takes with you, from finding you to coming back, with two numbers on each step. It takes one afternoon.

It borrows its rules from the Critical Client Flow in David Jenyns's SYSTEMology. Map one kind of customer and one service. Use two or three words per step, and list only what you actually do today.

Here are the seven steps for most service shops. Found, first contact, booked, showed, next job written, reviewed, came back. Found is reach, not a handoff, so you score the six steps after it.

For each step, write two numbers for last month. How many arrived at it? How many moved on?

Pull every number from a record. Your phone provider shows missed calls by hour. Your calendar shows bookings and no-shows, and your job or practice software shows the quotes and treatment plans written.

Why records and not memory? You remember the ad you paid for. You forget the quote nobody sent.

Fix the biggest drop first, because one step sets the pace

Your shop can only book as fast as its slowest handoff. Think of a relay team. The fastest runners still lose the race on one dropped baton.

So speed up any other step and nothing changes downstream.

There is always one slowest step, and once you fix it, the bottleneck moves somewhere else. That is normal. It means the fix worked.

Grove gives a second rule. Build the flow around the most expensive step, and catch problems before it, where they are cheap.

At Intel the costly step in hiring was flying graduates to the plant. So phone screens came first, and more plant visits ended in offers.

In a service shop the expensive step is the truck roll, the consult chair or the attorney's hour. That is our reading of the rule. Put the cheap check right before it: two questions on the phone, a confirm text the day before.

And if the drop is in delivery, more leads make it worse. One agency in the same SYSTEMology book went from three or four new clients a month to thirty, fell months behind, and had to stop its ads.

Once it mapped its flow and trained new hires from the map, it ran ads for a full year without a pause. First time ever.

More ads into a broken step is not growth. It is a faster way to let people down.

Most local businesses do not need more demand

The cleanest proof comes from a homebuilder, told in Demand-Side Sales 101 by Bob Moesta. When he joined the Detroit builder, it sold about a hundred houses a year from seven sites.

Every visitor got the same features pitch. Most stopped listening after about three minutes.

So the team interviewed recent buyers. Many were downsizers stuck on the move itself: the packing, the basement, the dining room table.

The builder raised its price and bundled movers, packing and a storage unit across the street. Sales rose 22 percent, by the author's account.

Then it moved its ads from the real estate pages to the obituaries, where buyers were already deciding. Traffic rose 37 percent and ad costs fell 70 percent.

Three questions about where each buyer stood cut a twenty-minute pitch to two minutes. The builder grew to fourteen sites and three hundred homes a year.

What did not change? The houses.

Every fix was a handoff: the pitch, the move, the moment of deciding. The homeowner who calls at 9 p.m. with no heat is deciding too. Do not drop that call.

Change the setup, not the reminder

Run every new step by hand from a written rule until it works, then automate it. That is SYSTEMology's rule, and it saves you from automating a mess.

Jenyns tested it on his own agency, where some accounts ran 120 days overdue. Three automated overdue emails barely moved anything.

Then the agency changed the setup: new clients paid up front by automatic billing before work began. Payments arrived in advance, on time, every time.

Michael Gerber's team in The E-Myth Revisited fought blue marker on white walls for more than three weeks. Meetings, memos, signs and spot checks all failed.

A clear plastic collar, four inches wide, around each board ended it at once. Engineers call that mistake-proofing.

Most small-shop automation is a memo campaign with a software bill. A reminder text is a memo. A rule that holds a slot only after the customer confirms is a collar.

So pick the first system from your biggest drop. Answer every inquiry in minutes, as speed to lead lays out, with an AI receptionist for after hours.

Or make every finished job write the next one. Or ask for every review and answer every one.

Write down what customers tell you, once

The best system in a small business is memory that never forgets. Gerber tells of a small hotel that asked his coffee brand and his newspaper, once.

Every visit after that, both were waiting. Nobody asked again. His words: "somebody had heard me."

Your shop hears things like that all day. The gate code, the dog's name, the filter size, the patient who hates needles.

Most of it lives in one person's head, and leaves when she does. Write it into the customer record and put it on the next job ticket.

Reviews work the same way. Read them weekly for the words customers use, the small version of a customer truth system. Then reply to every one, with a person sending it.

Is that marketing? It is the reason they come back.

What it costs, and what to skip

The honest cost is mostly time, not software. The first build is usually a rule inside tools you already pay for.

So price the gap, not the tool. Here is the sum with illustrative numbers.

Say your map shows 30 slow or missed inquiries a month, one in four would have booked, and a first job is worth $400. That is $3,000 a month going to someone else. If catching them costs less, build it.

Then count the owner's time. Plan on 30 minutes every Monday, plus an hour a week in the first month to correct drafts. A system you will not read is one you should not buy.

How to automate a 20-person shop into chaos in four easy steps. Step one: buy the tool from the conference demo.

Step two: connect it to everything on day one. Step three: let it text customers while nobody reads the drafts.

Step four: blame the tool and buy another damn one. Four half-built systems are four new ways to drop a customer.

Skip three things outright. AI that calls strangers without consent, tracking pixels on patient pages, and anything that games reviews. The rules sit in the dated box below, because rules change.

Read one number every Monday

A Monday number is one count you read every week that shows whether the system you built is working. Restaurants do this better than most offices.

In Lean Analytics, the Solare restaurant's staff report one ratio every day: yesterday's staff cost as a share of revenue. About 24 percent is right; over 30 percent is bad.

The authors like it because it is simple, immediate, actionable, comparable and tied to how the business makes money. The owner also counts reservations at 5 p.m. and multiplies by five to staff the night.

Your shop's version is short. Inquiries, booked, showed and sold, by source, for last week. Pair each count with a quality check, like reviews answered or quotes followed up.

Why pair them? A count on its own gets chased at any price. Grove pairs each volume number with a quality one, so the second shows what the first one costs.

Read it for four weeks against last month before you start a second system. If it does not move, change the rule, not the tool.

Then send booked jobs back to your ad platforms, so they learn from buyers instead of form fills. Server-side conversion tracking covers the how, and the full weekly memo is the growth cockpit.

Run the Handoff Map with AI this week

The prompt below runs the whole method. Paste it into your preferred AI, answer five questions from last month's records, and it names your biggest drop and the first rule to build.

1. Pull last month's call log, calendar and invoices. Fill the seven steps.

2. Run the prompt. Write the one rule on paper and run it by hand for a week.

3. Set a check date four weeks out. Read the number every Monday until then.

Want to score the rest of your systems? The marketing engineering audit does it in 20 questions.

Rather have the first one built for you? That is our marketing engineering work.

Remember the restaurant? One missing sentence on one call cost the whole dining room its wine for a night.

Most shops do not need more water poured into the bucket. They need to find the hole.