How to Build a Touch Calendar for Your Customer Base

Why this matters

Most shops contact their customer list when work gets slow, which is the worst possible timing: the touch goes out in a panic, the customers who respond arrive in a week you are already scrambling, and nothing happens again for eight months. A touch calendar replaces that with a plan you build once and run for a year. The thing that makes it work is not the copywriting. It is that the calendar is fitted to your capacity to serve what comes back, and that is a constraint almost nobody calculates before they start sending.

Step 1: Segment the list into exactly three classes

More than three classes is a system nobody maintains. Use behavior from your own records, not intuition.

• Class A: your base. Agreement holders, plus anyone with three or more tickets in the last 36 months. These earn real, personal contact.
• Class B: the middle. One or two tickets in 36 months. Most of your list lives here, and it is where the calendar has the most to gain, because these are customers who have not yet decided whether you are their shop.
• Class C: dormant. No ticket in 36 months, no known rupture. Low effort by design.

Pull the counts before you decide anything else. The shape of your list determines what is possible, and owners are routinely wrong about it by a wide margin in the direction of thinking Class A is bigger than it is.

Step 2: Set touches per class, and total them

A defensible starting point, to be tuned to your trade's natural service cycle: Class A four touches a year, Class B two, Class C one. Multiply through and get a single number: total touches a year. That number is the calendar's size, and everything after this is checking whether you can carry it.

Resist the urge to raise Class B to four. Doubling the touch count on the largest segment is the most common way a calendar becomes unrunnable in month five.

Step 3: Cost it in office hours, and expect this check to pass

Price each touch by channel in minutes. Reasonable planning figures: a merged, personalized batch touch costs roughly half a minute per record of office time once you amortize list preparation, review, and bounce handling. A phone touch costs roughly 4 minutes each including no-answers and callbacks.

Then add the part shops forget, which is usually larger than the sending: response handling. A response that turns into a conversation and a scheduled job costs on the order of 9 minutes of office time. Multiply your expected responses by that.

This check almost always passes with room to spare, and saying so plainly matters, because "we do not have time" is the standard reason a touch calendar never gets built and it is usually not true. The real constraint is the next step.

Step 4: Forecast the bookings, because that is what actually binds

Convert touches into expected work. Planning figures to start from and then replace with your own: a batch touch to a warm list responds in the high single digits as a percentage, and roughly half of responses become bookings. A phone touch reaches a real conversation maybe a third of the time, and a meaningful share of those book.

Now compare the resulting bookings against your service capacity in the weeks they will land. This is the number that decides whether the calendar is a plan or a problem. A calendar that generates work you cannot schedule for three weeks converts your own outreach into a disappointment, and it does that specifically to the customers you were trying to strengthen a relationship with.

Step 5: Load-level against your own seasonality, which means the opposite of instinct

Here is the rule most shops get backwards: schedule your heaviest touch batches 3 to 4 weeks ahead of your softest weeks, not ahead of your busiest ones.

Instinct says market before the season, because that is when customers are thinking about it and response rates are highest. That instinct optimizes response rate and ignores the fact that you did not need the work in those weeks. A touch that produces a booking in a week you were already full has not grown anything, it has displaced a job you would have had anyway, and it has spent a customer's willingness to respond on nothing.

Work backwards from your own job log. Find your four or five softest weeks. Put the big batches in the calendar at the lead time your booking pattern requires, which is usually 3 to 4 weeks out. Fill the rest of the year at a flat, low rate.

The pre-season instinct is not entirely wrong, and the exception is narrow: touches that are genuinely time-bound, like a warranty expiry or a real service interval, go when they are due regardless of your load. Those are information, not marketing, and delaying them for load-leveling is a disservice.

Step 6: Give every touch a real reason and a named channel

Every slot on the calendar carries three fields before it is finished: what it is about, which channel it goes on, and who it goes to. If you cannot write the reason in one honest sentence, the slot is empty and should be deleted rather than filled with something generic. A calendar with nine real touches beats one with twelve where three are filler, because the filler teaches the list to ignore you and the cost lands on the nine.

Channel follows consent and preference, not convenience. Text has its own consent rules and its own expectations, and running a marketing batch over text without explicit permission creates a problem that outweighs any response rate.

Step 7: Put it on one page

Twelve rows, one per month. Each row: which class, how many records, what the touch is about, which channel, and who owns it. It goes on a wall, not in a folder. A touch calendar that lives inside somebody's software as a set of scheduled tasks is invisible, and invisible plans are the ones that quietly stop running when the person who built them is on holiday.

Step 8: Name an owner and a fallback

One person owns the calendar. One named other person can run it. This exists because the failure mode of every touch program is a staffing change, and an obligation that lives in a single person's routine has a hard expiry date on their last day.

Step 9: Instrument it with two numbers per touch

Per touch: responses, and bookings. Not opens, not clicks. Write them on the same one page. At the end of the year you will be able to see which touches earned their slot and which were habit, and you will be able to cut the habits rather than cutting the whole program when it feels like too much work.

Worked example: a two-truck shop with 620 records

The list. 620 records total, segmented as: Class A 84, Class B 268, Class C 268. So Class A is about 14% of the 620, and the owner had been assuming it was closer to a third.

Touch count. Class A at 4 touches a year is 336. Class B at 2 is 536. Class C at 1 is 268. Total 1,140 touches a year.

Channel plan. Class A gets 2 phone and 2 batch, so 168 phone touches and 168 batch touches. Class B and C are batch only, 536 plus 268. Total 168 phone touches and 972 batch touches.

Office hours. The 972 batch touches at half a minute each is 486 minutes, about 8 hours a year. The 168 phone touches at 4 minutes each is 672 minutes, about 11 hours a year. Response handling: at a high-single-digit response rate the 972 batch touches produce roughly 78 responses, and at 9 minutes each that is 702 minutes, about 12 hours. Phone conversations that turn into scheduling add roughly 4 hours. Total about 35 office hours a year.

The office person had 4 hours a week available for this across 46 working weeks, which is 184 hours a year. The whole calendar consumes about 19% of the hours already set aside for it. The hours objection was not real, and confirming that in one afternoon of arithmetic is what got the program approved.

Bookings, which is the real constraint. From the 78 batch responses at roughly half converting, about 39 bookings. From the phone side, 168 calls reaching a conversation about a third of the time is roughly 56 conversations, and a meaningful share of those booking gives about 24 bookings. Total roughly 63 bookings a year from the calendar.

The shop ran 2 trucks at about 7 jobs a week each, 14 jobs a week, roughly 640 jobs a year. So the calendar's 63 bookings represent about 10% of annual job volume, which is a real contribution and comfortably absorbable across a full year.

The concentration problem. The owner's first draft put the big batches in March and September, ahead of both seasons. That draft concentrated most of the 63 bookings into roughly 8 weeks. Even taking half of them, call it 32 bookings across 8 weeks, that is about 4 extra jobs a week on top of a 14-job week, roughly a 30% load increase in weeks that were already the busiest of the year. On the actual draft, with more than half the bookings landing there, the surge was worse. That draft would have produced long lead times, pushed appointments, and a set of newly-contacted customers whose first experience of the outreach was being told to wait.

The load-leveled version. The shop pulled its softest weeks out of the prior year's job log: late January, most of June, and the first half of November. The big batches moved to 3 to 4 weeks ahead of those windows. The rest of the 1,140 touches spread flat. Target became roughly 5 to 6 bookings a month from the calendar, which against about 53 jobs a month of capacity (640 a year over 12 months) is under 10% of monthly capacity - absorbable in any month, including the busy ones.

Two touches stayed on their original dates and were deliberately exempted: a warranty-expiry notice and an equipment service-interval reminder, both of which are date-bound facts rather than marketing and both of which land in season.

Year-one instrumentation. The two numbers per touch showed something the owner had not expected: one of the Class B batch touches, a generic seasonal reminder, produced 3 responses out of 268 sent, about 1%, against the program's overall batch response in the high single digits. It was cut in year two and its slot given to a second Class A phone touch. That is the whole value of instrumenting - it lets you cut one slot rather than concluding the program does not work.

What changes the answer

If your trade has a long service cycle, three or more years between normal calls, cut every touch count roughly in half and lengthen the Class definitions to match. A two-touch-a-year cadence on a customer who needs you once every four years is a nuisance regardless of how good the message is.

If you are at or near capacity all year, do not build this yet. The correct move is pricing or hiring, and a touch calendar run by a shop with no soft weeks is a machine for generating disappointed customers. Build the calendar during the planning for the season after you have added capacity.

If your list is under about 150 records, skip the segmentation and the arithmetic. Work the list by hand, account by account, on what you actually know about each one. The structure in this article earns its keep somewhere in the low hundreds, and below that it is overhead standing in for judgment.

If a large share of your work is commercial, split those onto their own calendar. Commercial touches key off budget cycles, contract dates, and manager turnover rather than seasons, and mixing them into a residential cadence produces a calendar that serves neither.

How to verify you got this right

Three checks, at the end of the first year.

Look at the softest weeks from your prior job log and ask whether they were still the softest weeks. If they were, the load-leveling did not work and the touches either did not land or did not convert. That is the primary test of the whole design.

Count how many touches went out on their planned date. Under about 70% on time means the calendar is being run on remaining energy rather than as scheduled work, and the fix is the owner and fallback in Step 8, not more discipline.

Read the two numbers per touch and cut anything that materially underperformed the rest, then give the freed slot to the class where response was strongest. A calendar that is identical in year two to year one has not been instrumented, it has been photocopied.

References

• U.S. Small Business Administration (SBA), small-business marketing planning guidance
• Federal Communications Commission and Federal Trade Commission rules governing consent for commercial text and calling, which vary by channel and by how the contact was obtained
• See related: The Follow-Up Automation Worth Setting Up Once, The Customer Text Campaign Done Right, When to Stop Chasing a Dormant Customer, The Customer Anniversary Touch That Actually Works