Why Your Clients Say the Leads Didn't Work

You delivered every lead you promised, on budget, four months running. He canceled anyway and told two people at his chamber breakfast that your leads were junk.

It's 4:40 on a Thursday and his name is on your screen for the third time this week. You already know. He opens with "we need to talk about next month" and you hear a truck door slam behind him. He says the leads didn't work. You've got the dashboard open in another tab, 214 leads in four months, cost per lead down almost a third since March, and none of it will matter in the next ninety seconds. So you thank him for being straight with you, and then you sit there with a report nobody is ever going to read.

This page is about that call. Why it keeps happening to agencies doing the media work well, how to prove what actually went wrong with numbers you can pull this week, the exact conversation to have with a client who's halfway out the door, and what it takes to own what happens after the lead lands.

The leads were fine. The first hour wasn't.

Follow one lead through and you'll see it. A woman fills in a form at 8:12 on a Tuesday night because her water heater is leaking onto the laundry room floor. It lands in a group text his office manager checks in the morning. He's on a job at 8:12. He sees it at 9:40, calls once, gets voicemail, moves on. She booked someone else before breakfast.

Forty-two hours. Not because anyone's lazy. Because answering leads fast is a job, and in most of your clients' businesses it isn't anybody's job.

Read that with his account in mind. He isn't losing to your lead quality. He's losing twenty-five minutes at a time, over and over, for four months. Then he calls you.

You're being graded on something you don't touch

Here's the genuinely unfair part, and I'll say it once. You sold traffic and lead volume. You're judged on booked jobs. Between the two sit his phone, his voicemail greeting, the front desk person who quit in June, and his own willingness to call a stranger back at 8:15 at night. You control none of it and you get the invoice conversation anyway.

Being right about that is worth nothing by itself. He isn't running attribution. He's got one number in his head, what he spent against what showed up, and when it looks bad the simplest story wins. The simplest story is "the leads were bad."

You don't win by arguing. You win by changing what you both look at, and that takes four numbers he's never seen.

Four numbers that end the argument

Nothing about impressions, nothing about cost per lead. Each of these describes what happened after the lead left you.

1. Time to first contact. Clock starts when the form submits, stops at the first real attempt. Report the median. One lead answered at 3am will flatter an ugly month.
2. Contact rate. The share of leads that reached a live conversation. Not "we called it." Reached.
3. Follow-up attempts per lead. Attempts made before somebody stopped trying. If your median is one or two, you've found the problem.
4. Booked rate, split by speed. Appointments as a share of leads, cut two ways: contacted inside five minutes, and everything else. This is the one that ends it.

Put that next to a median of one or two attempts and you can see how much of what he paid for got thrown out without a fight.

Where to get the numbers when you don't own the CRM

The awkward thing about all four is that they live on his side. That's easier to solve than most owners assume.

• Your form or ad platform gives you the submit timestamp. That's your start time, and the only piece you own outright.
• Ask for view-only CRM access as part of reporting, not as an audit. Frame it as "so I can show you which sources actually convert."
• Call logs are the richest source and the most ignored. Most phone systems export outbound records with timestamps. Match them to the lead numbers.
• If he runs a shared inbox, the reply timestamp is your first-contact time. Rough, but honest and free.

You won't get clean data on the first pass. Get it directionally true, say so when you present it, and invite him to correct it. A number he's allowed to argue with is one he'll engage with.

How to say it without sounding defensive

The instinct is to lead with the defense. Don't. Open with "the leads were good" and it becomes a fight about your work, where he has to be wrong for you to be right.

• Open by putting your own money on the line. Say you'd rather find out now if you're wasting his budget.
• Show the split before the conclusion. Fast versus slow, same source, same ad. Let him do the math.
• Never name the person who dropped it. Blame the setup, not the human, or you'll lose the account winning an argument you'd won.
• Bring one change he can make this week. Not a strategy deck. One.
• Ask to change the contract number to booked appointments. That's the actual ask.

The version that works sounds like a partner reviewing a shared problem. The version that fails sounds like a vendor holding up a receipt.

The conversation, close to word for word

Steal this and change the wording until it sounds like you. Use it on a call, never email. Email lets him skim to the part he'd already decided.

Adding a response layer to what you already sell

Say he agrees the follow-up is where it leaks. You can hand him a fix he almost certainly won't implement, or you can take the leak off him. That second option is all a conversion layer is. You keep selling lead generation exactly as you sell it, and you add the part that answers the lead: an instant first response the moment a form hits, a voice that picks up an inbound call in under a minute and books into his calendar, then human follow-up on a real cadence.

GoStartr is built for this. It sits over whatever CRM your client already runs, no migration. Abby answers inbound leads in under a minute, qualifies, books the meeting and writes it back to his system. Calls get recorded and scored so you can show him coaching instead of a count. Your people still do the outbound calling. The AI only handles people who raised their hand first.

What it changes for you is retention, and retention is the business. Once your reporting moves from leads delivered to appointments booked, you stop being a line item he compares against a cheaper agency. Month four stops being a cliff.

What this will actually demand of you

The honest part, because most of the cost here isn't money.

• Somebody has to own it. Not "we all keep an eye on it." A name, a number, and a daily look at every lead that never reached a human.
• Your clients have to answer their phones. If he won't take a booked appointment at 4pm on a Wednesday, no layer saves him.
• It is not set and forget. Offers change, and a qualifying question that worked in March will be wrong by August.
• You own outcomes now. That's the point and the risk. When booked rate drops you get the call.
• Somebody has to listen to the recordings, or you've bought an expensive answering machine.

On commercials, I won't invent a number for you. Partner terms get set per partner, and what suits a five-client shop looks nothing like a forty-client one. That's a direct conversation, not a price list.

Who should not do this

This is the wrong move for plenty of agencies, and better to find that out here than three months in.

• If you've no appetite for client operations. This puts you inside his working day: escalations, no-shows, a rep who was short with somebody on a Tuesday. If that reads like hell, keep selling media and get better at setting expectations up front.
• If his sales process is the real problem. Plenty of businesses answer fast and still don't close, because the pricing is wrong or the quote takes nine days. Speed just gets him to the same no, sooner.
• If your leads genuinely are weak. Run the split first. If booked rate is flat between fast and slow contact, a layer only makes you accountable for a bad offer.
• If you're already at capacity. Adding operations to a team that's late on reporting loses you the clients you've got.

It fits best where you sell to service businesses, where the buyer compares two or three vendors on the same evening and gives the job to whoever picks up.

Do this before Friday, even if you never buy anything

One task. Pull the time-to-first-contact number for your three biggest clients.

1. Pick the three accounts it would hurt to lose.
2. Export the last thirty days of leads for each, with submit timestamps.
3. Get the first-contact timestamp however you can: call log, CRM activity, his sent folder.
4. Work out the median gap and the share contacted inside five minutes.
5. Split booked rate into fast contact versus everything else.

Inside two hours you'll know which of your three biggest accounts is going to cancel in month four, while there's still time to do something about it. Worth an afternoon whether you ever add a layer or not.