Which Leads to Call First: A Queue Order for Insurance Agents

Ask most agents how they decide which lead to dial next and you get one of two answers. Either “top of the list” or “whichever one I feel like.” Both are a system. Neither is a good one.

I dialed top-of-the-list for years. The list was sorted by whatever order the CSV came in, which usually meant alphabetical by last name or the order the vendor's form captured them. So I spent my best hour every morning on whoever happened to be named Anderson.

The fix is not complicated, and it does not need a scoring model. It needs a fixed order of tiers, applied the same way every session, so the decision of who to call next is already made before you sit down. Here is the order I use and why each tier sits where it does.

Why queue order matters more than dial volume

A lead's value is not fixed. It moves with time and with what the prospect last did. A callback that is due right now is worth more than the same callback two hours from now. A lead that filled out a form this morning is worth more than one that filled it out last Tuesday. An order that ignores this spends your sharpest, most energetic dials on records that were not going anywhere and reaches the time-sensitive ones after they have cooled.

The other reason is consistency. Dialing sessions degrade in the back half. When you are tired and picking leads by feel, you gravitate toward the comfortable ones — people who were friendly last time, records with lots of notes. Those are often the least likely to move. A fixed order removes that choice from the tired version of you.

This is the piece that sits before the session structure itself. The block tells you when and how long to dial. The queue order tells you who.

The four tiers, in order

Every record in the queue falls into one of these. The tier decides position; within a tier, sort by recency. I will explain each one, but the table is the whole system.

Tier 1: callbacks that are due

This is the only tier where the prospect made a commitment. They said “call me Thursday after two,” and it is Thursday after two. Everything else in the queue is you hoping for attention. This one is attention already granted.

It goes first because it is the only tier that gets worse if you delay it by an hour. A new lead you reach at 11 instead of 10 is roughly the same lead. A callback you make at 3:30 instead of the 2:00 you promised is a different conversation, and usually a shorter one. The return-call playbook covers what to say; the queue's job is just to make sure you say it on time.

One rule: a callback only belongs in tier 1 if the time was set with the prospect. “I'll try them again next week” is a note to yourself, not a callback. Those go in tier 3.

Tier 2: fresh leads and inbound contact

Two kinds of record live here. Leads that came in during the last 24 hours, and anyone who reached out to you — a missed call from a number in your CRM, a text reply, a voicemail.

Fresh leads are here because of speed to lead. Whatever the exact curve is for your lead source, it slopes down, and the steepest part is the first day. A lead that is 20 minutes old is the one exception to tier 1 going first: if a callback is due at 2:00 and a lead drops at 1:52, I take the lead, because eight minutes on a callback is a rounding error and eight minutes on a brand-new lead is not.

Inbound contact is here because the prospect did the hardest part for you. Someone who called back a missed call has already decided you are worth a conversation. The number of agents who let those sit until “after I finish my list” is a big reason contact rates on inbound are as low as they are.

Tier 3: in-progress

These answered at least once, showed some interest, and did not set a specific callback. The “send me something” people. The “let me talk to my daughter” people. The ones who were interested but got interrupted.

They go above the untouched pile because you know two things: the number is live and connects to the person on the record, and they did not hang up on you. That is more than you know about any tier 4 record. The right sort here is longest-since-last-touch first, so nobody quietly ages out of the pipeline because they were never at the top.

This tier is also where pipeline rot lives. If a record has been in tier 3 for 60 days with no answer on the last five attempts, it is not in progress anymore. Disposition it honestly and let it fall to tier 4 or out of the queue entirely. The monthly pipeline audit exists to catch exactly these.

Tier 4: untouched and aged

Everything else. Leads older than a day that never answered, imported lists, records that came back from a revival campaign. This is where most of your dials happen, and it is also where the order matters least, because the value of any one record is roughly the same as the next.

Sort by fewest attempts first. A record you have dialed once has more upside left than one you have dialed six times, and working the low-attempt records first spreads your touches evenly across the list instead of hammering the same two hundred names. Once a record hits your cadence limit — whatever number you settled on in your cadence plan — it leaves the queue.

What gets filtered out before any of this

Ordering only applies to records you are allowed to dial. Before the tiers, a few filters remove records from the queue entirely, and these are not optional.

• DNC. Anything on the national registry without a valid exemption, anything on your internal do-not-call list, and anything the prospect told you to stop calling. These are not tier 5. They are not in the queue.
• Outside calling hours.A record in a time zone where it is currently before 8 a.m. or after 9 p.m. local — or tighter, in states with their own windows — is invisible until the window opens. A good queue re-surfaces it automatically when it does.
• Consent expired or missing. If your consent standard for that lead source has a shelf life and the record is past it, it is out until you have a new basis to call.
• Cadence exhausted. Hit the attempt cap, no answer. Out of the active queue and into whatever your long-term nurture looks like.

I put the filters after the tiers in this article because the tiers are the interesting part. In the software they run first, every time, before you see a single name. That is the point of having them in the software.

Building the order in your CRM

None of this needs a special feature. It needs four fields filled in consistently and a saved view or two.

1. Next-action date.Every record that is not dead has one. For a callback it is the agreed time. For everything else it is when the cadence says to try again. Tier 1 is simply “next-action date is now or earlier, and the disposition is Callback Set.”
2. Created date.Already there in any CRM. Tier 2 is “created in the last 24 hours and attempt count is zero,” plus anything flagged inbound.
3. Last-contact date and disposition.Tier 3 is “last disposition was a live answer that was not a close or a refusal, and no callback set.” This only works if your disposition codes distinguish a live conversation from a voicemail, which is the single most common thing agents skip.
4. Attempt count. Incremented on every dial, automatically if your dialer talks to your CRM. Tier 4 sorts on it ascending.

In a spreadsheet you can fake this with a formula column that outputs a tier number and a sort. It works for a few hundred records and then stops being maintained, which is roughly the point where a sheet stops working in general. In a CRM with a built-in dialer, the queue is the view, and the dialer just walks it top to bottom.

Where lead scoring fits, and where it does not

Agents ask about scoring because vendors sell it. The honest answer for a solo agent or a small team is that it rarely beats the four tiers, and it often does worse.

Scoring models need history to learn from. A thousand closed records with clean dispositions is a starting point. Most independent agents have a few hundred records and dispositions that were entered in a hurry. A model trained on that learns your habits, not your prospects. If you historically called the friendly ones more, the model will tell you to call the friendly ones more.

The tier system is not a model. It is three facts about each record — did they commit, how recent are they, have they ever answered — and those facts are true regardless of your history. It captures most of what scoring would give you without the setup, and it is easy to explain to a new agent on their first day.

Scoring earns its place when an agency has thousands of shared records, several dialers pulling from the same pool, and someone whose job is to maintain the data. Below that, skip it.

A session, in practice

Here is what the order looks like on a normal morning block. Two callbacks due between 9 and 10, so those are the first two dials at their set times. Between them, three leads dropped overnight from a form, so those fill the gaps — newest first. Then a text reply from yesterday that I had not answered. That is tier 2 done, and it took maybe twenty minutes.

Tier 3 next: fourteen records that answered in the last month and went quiet, oldest touch first. Some of those become tier 1 callbacks for later in the week. A couple get dispositioned out. Then the rest of the block is tier 4, lowest attempt count first, and I stop looking at the list because there is nothing to decide anymore. Dial, disposition, next.

The part that changed my numbers was not the tier 4 volume. It was that tier 1 and tier 2 never got buried under it again.