Conversion rates almost always fall when volume rises, and the cause is nearly never the leads. Where the leak opens up, and the order to fix things in.
Figures in this article describe the wider market and are drawn from the third-party sources listed at the end. They are not Lead Foundry results, and nothing here is a projection of what any individual broker will achieve.
A broker doubles their lead order and their conversion rate falls by a third. The supplier is the obvious suspect and is usually innocent. What changed is that the process which comfortably handled ten records a week is now handling twenty, and processes fail at their weakest point rather than evenly.
This is where the leak opens, how to find it from your own data, and the order to fix things in.
Because attention is not divisible. When a broker with capacity for ten leads receives twenty, they do not give each one half the attention. They work whichever records happen to arrive when they are free, at full effort, and give the rest a single call whenever the diary next opens.
Given how sharply contact odds decay, that second group performs terribly. Averaged together, the batch looks like a quality problem. Split by time-to-first-attempt, it resolves into two populations, one performing normally and one barely worked.
The diagnostic that settles it in ten minutes: Split last month's leads into those contacted within an hour and those contacted after four hours, then compare contact rates. If the gap is large, you have a capacity problem. If both groups are equally poor, you have a supply problem worth raising with your supplier.
In a predictable order, because each stage protects the one after it.
| Stage | What breaks under volume | The symptom in your numbers |
|---|---|---|
| Delivery and assignment | Leads pile up unowned in a shared inbox or queue | Time to first attempt rises and its variance widens |
| First contact | Attempts slide from minutes to hours | Contact rate falls while data quality is unchanged |
| Cadence adherence | Attempts per lead fall as the queue grows | Average attempts drops toward two |
| Appointment conversion | Calls get rushed and shallower | Contact rate holds but bookings fall |
| Appointment delivery | The diary has no room for the meetings booked | Bookings hold but meetings are scheduled weeks out |
The last row catches firms that fixed everything upstream. There is no point improving your appointment rate if the appointments land three weeks away, because the drop-off between booking and attending grows with the gap.
Upstream first, always, because a fix downstream is measured through a broken stage upstream and will look like it did not work.
The evidence for the third item is direct: research across close to 3.5 million leads found that 93% of converted leads are reached by the sixth attempt and that 50% of leads are never called a second time. Salesforce reports 44% of reps stop after a single attempt while 80% of sales require five or more follow-ups. Cadence adherence is where the largest recoverable loss usually sits.
| Converted leads reached by the 6th call | 93% |
|---|---|
| Leads never called a second time | 50% |
| Leads that never receive one email | 59% |
| Reps who stop after a single attempt | 44% |
Three of these four bars describe effort that was paid for and not spent. Volume increases make each of them worse, which is why capacity has to be added before volume rather than after.
Source: Velocify contact-strategy research and Salesforce State of Sales
Capacity is either hours or automation, and most firms need some of each. The sequencing rule is that the capacity has to exist and be proven for a month at current volume before the order size changes.
| Lever | What it buys back | When it makes sense |
|---|---|---|
| Automated assignment and task creation | The minutes between arrival and ownership | Immediately, at any volume |
| Protected calling blocks | Reliable first-attempt timing | Immediately, at any volume |
| An administrator handling file work | Adviser hours, since roughly 60% of sales time goes to non-selling work | Once volume is consistently at capacity |
| A second adviser or a dedicated caller | Genuine additional calling capacity | Once the process is documented enough to hand over |
The third row is where the Salesforce figure earns its place: sales professionals spend around 40% of their time actively selling. For a broker, the non-selling 60% is compliance files, applications and administration, and it is the cheapest capacity to buy back because it does not require another licensed person.
Increase in increments you can reverse, and hold each increment long enough to read the result. A step of 25% held for a full month gives you a clean comparison. Doubling the order and changing your CRM in the same fortnight gives you nothing you can interpret.
Almost always because time to first contact has lengthened. Extra volume sorts by arrival time rather than spreading evenly across your attention, so records arriving while you are busy get one late call. Split last month's leads by how quickly they were first attempted and compare contact rates. A large gap between the fast and slow groups points at capacity, not at the supplier.
Add capacity first and prove it holds for a month at your current volume, then raise the order by no more than 25% and hold that level for a full month before judging it. Watch median time to first attempt weekly, because it degrades before revenue does and gives you time to reverse the step.
Before, and usually the first hire is administrative rather than another adviser. Sales professionals spend roughly 40% of their time actively selling, so buying back the compliance and application work returns adviser hours without needing a second licensed person or a longer handover.
Delivery and assignment. If leads sit unowned in a queue or an inbox while somebody notices them, every downstream improvement is measured through that delay and will look ineffective. Automate assignment so a lead has a named owner and a task the moment it arrives, then work on speed, then cadence, then the script.