Kept Plumbing Call Study

What 1,394 inbound calls to a Long Island plumbing company reveal about when homeowners actually pick up the phone — and which of those calls turn into work.

The headline numbers

What is the Kept Plumbing Call Study?

The Kept Plumbing Call Study is an analysis of every inbound phone call received by a single owner-operated plumbing company on Long Island, New York, over 141 consecutive days between March 13 and July 31, 2026. The business routes 100% of its inbound calls through Kept's AI receptionist, so the dataset captures every call the business received in that window — not a sample, and not only the calls someone had time to pick up.

We publish it because nobody else is publishing it. There is a great deal of survey research asking homeowners what they say they do, and very little call-level data showing what they actually do. This is one business, in one trade, in one county. It is not the industry. It is a complete record of one phone line, and it is more than we could find anywhere else.

How the data was collected and classified

Every call was logged automatically by the voice platform handling the business's line, then deduplicated by unique Call ID. No calls were sampled, excluded for convenience, or added by hand.

Each call was classified from its transcript along two dimensions: when it arrived, and whether it represented real work.

Classification was performed by Kept from the full call transcripts. The underlying business is not named at its owner's preference; it is a licensed plumbing and heating contractor operating in Suffolk County, New York.

What does "after hours" actually mean?

"After hours" is usually shorthand for "evenings and weekends." In this dataset that shorthand is wrong twice over. The 397 calls arriving outside business hours split across three windows: weekday evenings from 5:00 PM onward accounted for 40.1%, weekday mornings before 9:00 AM accounted for 31.7%, and weekends accounted for 28.2%. Nearly a third of after-hours calls came in before the workday started — a window the common phrasing leaves out entirely.

The lead rates make the point sharper. Weekday evenings are the largest after-hours window by volume but the weakest by quality, qualifying as leads at 39.6% — barely above the 37.4% business-hours rate. Early mornings ran 47.6% and weekends 64.3%. An owner who decides to cover "evenings and weekends" is naming the one after-hours window that behaves most like an ordinary weekday afternoon.

All 397 after-hours calls, March 13 – July 31, 2026. Percentages in the share column are of after-hours calls, not of total call volume. Weekend covers all hours on Saturday and Sunday; the two weekday windows cover Monday through Friday only.

Is the after-hours rate steady from month to month?

No. The blended figure across the full period is 28.5%, but the monthly rate ranged from 24.6% in April to 37.8% in May. Any single month can mislead in either direction. We publish every month separately so the variation is visible rather than hidden inside an average.

March is a partial month — service began March 13, 2026. Do not average the monthly percentages; the blended figure is calculated from total calls.

Are after-hours calls worth answering?

Yes, and by a wider margin than expected. Calls arriving outside business hours qualified as leads at 49.1%, compared with 37.4% during business hours — a difference of nearly twelve percentage points. After-hours calls in this dataset were not leftover junk that survived until evening. They were, on average, the better calls.

One plausible explanation is that a homeowner calling at 7:00 AM or 8:00 PM is calling because something is wrong right now, while a weekday-afternoon call is more likely to be routine. We can observe the difference; we cannot prove the cause from call data alone.

What share of weekend calls are real jobs?

In this dataset, 64.3% of weekend calls were qualified leads — 72 of 112 calls. That is the highest lead rate of any window we measured, and it is the most counterintuitive result in the study.

Weekends are a small slice of total volume, at 8% of all calls. But nearly two out of every three weekend calls represented real work. For a business that lets the phone go unanswered on Saturday and Sunday, the cost is not proportional to the call count.

The caveat matters: 112 calls is the smallest sample in this study, and a rate calculated on 112 calls will move more than one calculated on 1,394. We will republish this figure as the dataset grows.

What this study does not show

One business.

Every figure here describes a single plumbing and heating contractor. A second business, even on the same block, could produce a different distribution.

One trade.

Plumbing has an emergency profile that general contracting, remodeling, and most other trades do not share. Do not transfer these rates to a trade where nothing breaks at midnight.

One market.

Suffolk County, New York. Suburban, seasonal, and Northeastern. Call patterns are shaped by climate, housing stock, and commuting hours, none of which are constant across the country.

141 days, spring into summer.

This window does not include the winter heating season, which is the busiest period of the year for this business. The next update will begin to close that gap.

Calls, not outcomes.

We classified whether a call represented real work. We did not track which of those leads became booked jobs or collected revenue.

An assisted phone line.

Every call in this dataset was handled immediately, because the business routes all calls through an AI receptionist. This is a record of what happens when the phone gets picked up — not a measurement of what happens when it does not.

Findings you may cite

Each finding below is written to stand alone when quoted. If you are citing this study, please use the wording as written — it names the dataset, the period, and the sample size, which is what makes the claim checkable.

How to cite this study

This study is published under a Creative Commons Attribution 4.0 International license. You may quote, republish, and build on these figures — including commercially — as long as you credit Kept and link back to this page.

Kept Plumbing Call Study. Kept, Version 1.0, August 14, 2026. Data covering March 13 – July 31, 2026. https://getkept.ai/call-study

License terms: creativecommons.org/licenses/by/4.0/

Version history

This page is updated as the dataset grows. The March–July 2026 baseline above is fixed and will not change — later periods are added alongside it, so any citation of this page remains accurate.

• August 14, 2026 — Version 1.0

  First publication. Baseline dataset covers March 13 through July 31, 2026.