Most operators quote an average speed to lead and feel fine about it. The average is dominated by the leads that came in at 10am on a Tuesday while someone was sitting at a desk. It tells you nothing about the 8pm lead, which is the one you lost.
An average is a single number describing a distribution that is not remotely uniform. Your response time is not one number with a little variance around it. It's two completely different businesses stapled together: the one that runs while somebody is at a desk, and the one that runs while nobody is.
When you average those together you get a figure that describes neither. And because business-hours leads outnumber after-hours leads, the average sits close to the good half and quietly buries the bad half.
Export leads created and first outbound attempt for the last ninety days. Bucket by hour of creation. Look at the median response time in each bucket. The shape of that chart is the actual story, and for most shops it falls off a cliff after 6pm and never recovers until morning.
Two details matter here or the exercise gives you a comfortable answer instead of a true one.
One lead that sat for four days will drag a mean into nonsense. The median tells you what actually happens to a typical lead in that hour. If you want to be thorough, pull the 90th percentile too. That's your worst realistic case, and it's the one that's costing you.
This is where most operators accidentally flatter themselves. A dial at 8:04pm that rang out is not a response. Neither is a voicemail. If your report measures first attempt, you'll see a great number for a lead nobody ever spoke to.
Measure the gap between lead created and first two-way conversation. That's the number that correlates with whether you got the deal. Pull both if you can, and the gap between the two columns is its own finding.
Every shop is different, but the pattern repeats. Response times sit tight through the mid-morning and early afternoon, widen through the late afternoon as the team gets into appointments, then step off a cliff somewhere between 5pm and 7pm depending on when your phones stop.
Overnight, the median is not a response time at all. It's the distance to the next business morning. A lead at 9:15pm and a lead at 2:40am get the same answer, which is roughly 9am, and both look identical in a report that only tracks the average.
Weekends are the same story with a longer tail. A Friday evening lead is often a Monday morning lead. If your report bins by hour but not by day, run it again split by weekday and weekend. Most operators find their single worst window is Friday after 5pm, and it's also one of their highest-intent windows.
Under five minutes, in every bucket, including 2am. Not because 2am leads are common, but because the leads that arrive when nobody is watching are the ones your competitors also aren't answering.
That last clause is the whole opportunity. The 10am lead is contested. Everybody answers the 10am lead. The 8:40pm lead is contested by whoever happens to pick up, which is usually nobody, which means it's not really contested at all. Being the one shop that answers it is the cheapest edge available to you, and it requires no additional marketing spend.
Because a seller calling signs at night is not calling one sign. They're working down a list on their phone, and the ordering is whoever answers. If you call back in the morning, you're not competing on your pitch. You're competing against a contract somebody else already signed.
The window isn't about being impressive. It's about arriving while the decision is still open.
Once you have the chart, ask what percentage of your monthly lead volume lands in the bad buckets. This is the number that turns an interesting chart into a business case.
If 12 percent of your leads arrive after 6pm and on weekends, and those leads get a median response measured in hours instead of minutes, you now know precisely how much of your marketing spend is being routed into your weakest coverage. Multiply it out with the same math from the lead leak calculation and you have a dollar figure instead of a shape.
Rotate an on-call phone. Cheapest, and it works for about six weeks. Then whoever is holding the phone starts screening, because answering unknown numbers at 9pm on your own time is not sustainable, and the coverage quietly degrades back to where it started.
Hire an overnight setter or a VA in another time zone. This genuinely works and plenty of shops run it. It's a real payroll line, it needs managing, and coverage still has holidays, sick days and turnover.
Answer it automatically. No shift to fill and no drop-off, but only if the thing answering actually qualifies rather than just taking a message. A message taken at 8:40pm and read at 9am is the same lost lead with extra steps.
There's no free option in that list. The point of pulling the chart is that you get to pick with a number in front of you instead of a feeling.
Two things will quietly distort it, and both are common.
Marketplace and aggregator leads sometimes carry the timestamp of when they were generated, not when they hit your system. If a batch lands in your CRM at 6am carrying last night's creation times, your overnight bucket will look catastrophic and your morning bucket will look heroic, and neither is real.
Check one batch by hand. If created-at and received-at differ, split them into their own bucket or measure from received-at instead.
If an automatic text fires the instant a lead arrives, and your report counts that as the first attempt, every bucket will look excellent and the chart will be useless. Exclude automated touches, or track them in a separate column. The thing you're measuring is time to a human conversation.
Whatever you change, run the exact same export again a quarter later and lay the two charts side by side. This is the only honest way to know whether a change worked, because deal flow moves for a dozen reasons and you'll credit the wrong one otherwise.
Keep the methodology identical. Same ninety-day span, same buckets, same median, same definition of a conversation. If you change how you measure at the same time you change how you operate, you've learned nothing and you won't know it.
Pull ninety days. Bucket by hour. Use the median. Measure to first conversation. Look at where the line breaks, and then look at what share of your volume lands on the wrong side of it.
Most operators have never seen this chart for their own shop. It takes an afternoon to build and it tends to change what you do next.
A simple way for real estate operators to calculate what unanswered and unreturned leads are actually costing them each month.
Trace one after-hours seller lead end to end through your own system. Most operators find the lead dies in a place nobody owns and no report shows.