AI in Sales
They Didn't Pick the Cheapest Plumber. They Picked the One Who Picked Up.
The belief that plumbing customers always buy on price survives because the jobs you lose to silence never tell you why. What actually decides the call, five fixes you can start this week, and the planned-work cases where a lower number genuinely does win.
by Jerrod Anthraper
A homeowner watches water come through the kitchen ceiling and calls three plumbers off the first page of Google. One answers. Two go to voicemail.
She books the one who answered, at a price she never compared against anything, because there was nothing yet to compare it to. That evening she tells her husband she shopped around. She believes it. He believes it. Somewhere across town, two owners are looking at their close rate and concluding the market has gotten cheap.
The myth, stated plainly
The belief goes like this: plumbing is a commodity, homeowners collect three quotes, and the lowest number wins. Owners who believe it respond by shaving margin, running discount offers, and training their techs to lead with price.
The belief is wrong most of the time, and the reason it survives is that the evidence against it never reaches you.
Consider what the research on response timing says: 78% of sales go to whoever responds first. In a trade where the buying decision is frequently made under stress, with water on the floor, that number is not a marketing curiosity. It is a description of how your calls get won and lost before anyone talks about money.
Why the myth is so convincing
Because you only hear about the losses where price got said out loud.
The customer who tells you "I found someone cheaper" is a data point that walks up and introduces itself. The customer who never calls back because your phone rang eleven times is not a data point at all. She simply never enters your world, and the loss shows up as an absence in your schedule rather than a reason in your CRM.
Your team then compounds it honestly. CSRs and techs hear price objections from the minority of callers who actually shop on price, because those are the only callers who raise the topic. They report what they hear. Nobody reports the silence.
So the shop ends up optimizing against the loudest tenth of its losses and never touching the quiet majority.
There is a second reinforcing loop. Once an owner decides the market is price-driven, the discount offers go out, and discount offers attract exactly the callers who shop on price. The belief then confirms itself with data the owner generated. Six months in, the customer mix genuinely has gotten cheaper, and nobody remembers that the strategy came first.
What to fix instead
None of the following requires software or a new hire. All of it can be started this week.
1. Measure your answer rate, not your call volume
Call your own main line at 7:10 in the morning, at noon, at 6:40 in the evening, and at 8:00 on a Saturday. Log what actually happens each time: live human, hold, voicemail, or an answering service reading a script.
Then pull your phone records for last month and count inbound calls that lasted under twenty seconds. Those are almost entirely people who hung up. That count is your real problem list, and most owners have never once looked at it.
2. Separate "answered" from "booked"
Answering is not selling. Plenty of shops with good answer rates still lose the call, because the conversation ends in a promise instead of a slot.
Take fifty answered calls from last week and mark each one: booked on the call, promised a callback, or ended with nothing. If your callback bucket is large, you do not have a pricing problem or a staffing problem. You have a handoff problem, and handoffs are the cheapest thing in this business to fix.
Listen to five of those calls if you record them. The pattern is almost always the same: the caller was ready, the CSR did not have visibility into the schedule, and the conversation ended in a maybe because nobody in the room could say yes.
3. Give a real arrival window on first contact
"We will have someone call you back to schedule" is the single most expensive sentence in home services. It converts an interested buyer into a person who now has to wait, and waiting people keep dialing.
Replace it with a specific window on the first contact, even a wide one. "I can get a tech to you between two and five today, or first thing tomorrow morning" removes the reason to call the next company on the list. Certainty is the actual product she is shopping for. Price is just the number she uses to describe her uncertainty.
4. Price the diagnostic, not the job
Homeowners ask "how much" on the phone because it is the only question they know to ask. Refusing to answer teaches them you are hiding something. Guessing at a full job price teaches them to hold you to a number you cannot honor.
Quote the diagnostic instead, plainly and out loud, and say what it covers and what happens to that fee if they proceed. You have now answered the question, established that you are not evasive, and moved the conversation from an imaginary number to a real appointment.
5. Give the estimate a decision date
An estimate that ends in "let us know" ends in nothing. Every unsold estimate should carry a date attached to something real: parts availability, a scheduling slot held, the price validity window, or the seasonal calendar.
Then follow up on that date, once, in writing. Not to pressure. To close the loop, which most of your competitors will never do.
The honest exception
Sometimes it really is price. For planned work with no urgency attached, such as a water heater swap, a repipe, or a sewer line replacement scheduled for next month, homeowners do collect written quotes and do compare them line by line. In those situations a genuinely lower number wins jobs, and pretending otherwise is not useful.
The point is narrower than "price never matters." The point is that price gets blamed for losses it did not cause, and the misattribution sends owners cutting margin when the actual fix was answering the phone.
What the timing data shows
Data point: responding within 60 seconds can lift conversion by as much as 400%.
Sit with what that implies about the first-page competition. If three companies show up in her search and one responds in under a minute while the others respond in an hour, they are not competing on the same terms at all. One of them is having a conversation. The other two are leaving messages for somebody who has already booked.
The lowest bidder cannot win a job that was booked before the bid existed.
Where this goes from here
Steps one through five are all achievable with a disciplined team and a slow week. The trouble is that slow weeks are not when you lose the calls. You lose them on the day three techs are out, the dispatcher is covering two phones, and every homeowner in the county discovers their pipes at once.
Tykon's AI agent, James, is built for exactly that day. He answers inbound calls and form fills in under 60 to 90 seconds, around the clock, asks the qualifying questions your dispatcher would ask, gives the arrival window, and books the job into the real schedule. Afterward he handles the review request and the referral prompt that your techs will never remember to make from a crawlspace.
Start with the cheapest test in this article. This Saturday at eight in the morning, call your own main line and listen to what a homeowner hears. Whatever happens next is the honest answer to why you lost last month's jobs.