# How to Cut Lead Response Time in an Insurance Agency: Measurement, Routing and a Routine

> Not just “answer faster”: how to measure response time, route leads, handle after-hours leads, and build a weekly routine that shortens it. With a free calculation model.

- Published: 2026-10-25 · Updated: 2026-10-25
- Running an agency · 9 min read · The Financi team
- Canonical page: https://www.financi.co.il/en/articles/lead-response-time-insurance

## The short version

- You cannot shorten what you do not measure. Step one is knowing today's response time, by hour, day and source.
- A lead needs an owner the moment it arrives. Automatic routing solves the “who takes it?” question behind most delays.
- Most delay arises in the same places: out-of-hours, breaks, holidays, and leads arriving in several channels. Each has a planned answer.
- Automation raises speed but does not replace a person. An automatic confirmation is excellent; an automatic sale is not.
- A short weekly routine over an exceptions report improves more than a one-off “answer faster” campaign.

Almost every agent will agree that answering fast matters. Few know how long a lead actually waits with them. The article [speed to lead](https://www.financi.co.il/en/articles/speed-to-lead-agents) covers the first response itself and its scripts. Here we cover the system around it: **how to measure, how to route, what to do after hours, and how to build a routine that shortens the time over months.**

## Measure first

Without measurement every conversation about speed is an argument about feelings. Three timestamps per lead are enough, and in the first two weeks you can record them in a spreadsheet:

| Timestamp | What it is | What you calculate from it |
| --- | --- | --- |
| Lead arrives | The moment the details came in, by source (form, WhatsApp, phone, referral) | Time to assignment: how long until there is an owner |
| First response | The first human message or call, not an automatic confirmation | First response time: the main measure |
| First actual conversation | The first time you spoke with the client | Time to conversation: how fast the lead moved on |

Cut the data by hour of day, day of week and lead source, and look for outliers rather than the average. An average hides the worst cases, and most of the problems are in them. Do not set a target before you have seen your own agency's data. A number copied from another agency does not reflect your team, channels and hours.

## Routing: a lead with no owner goes cold

Most delays start with the question “who takes this?”. When an owner is set for every lead without human involvement, it disappears. Three common methods:

| Method | How it works | Advantage | Drawback |
| --- | --- | --- | --- |
| Queue (round-robin) | Leads are shared in turn among active staff | Simple and fair | Ignores specialism and load |
| By topic or source | Pension to one person, health insurance to another, by lead source | Matches lead to agent | Needs a rule for an enquiry that spans topics |
| By existing client | A known client's lead returns to their agent | Continuity and personal service | Needs a stand-in when the agent is unavailable |

> **Always define “what happens when they do not answer”.** Good routing includes a fallback rule: if the owner has not responded within a time you set, the lead goes to a stand-in and a manager is alerted. Without a fallback, routing just moves the delay to one person.

## Where the delay actually arises

| Situation | Why it delays | Planned answer |
| --- | --- | --- |
| After working hours | The lead waits until morning | An orderly confirmation saying when you will answer, and a morning queue that starts with the night's leads |
| Peak-hour load | The whole team is on calls | A queue that distributes, and an on-duty person for first response only |
| Holiday or sick leave | The lead stays with someone who is not there | A stand-in defined in advance and routing that skips whoever is absent |
| Several channels in parallel | The same lead arrives by form and by WhatsApp | Merge into one card so nobody answers twice |
| A lead that lands in an inbox nobody checks | There is no alert | An immediate alert to the owner and a daily exceptions check |

## Automation: what yes and what no

Good automation shortens the time before the person, not instead of the person. What suits automation: confirming receipt, assignment, alerting the owner, a reminder when time has passed, and asking for one missing detail. What does not: advice, a recommendation and a promise about a product. [The article on automating the admin](https://www.financi.co.il/en/articles/automate-the-admin-hours-saved) details what is worth handing to a machine. In Financi the shared inbox and the processes are tied to the lead and the card, as described in [the WhatsApp module](https://www.financi.co.il/en/whatsapp) and [the CRM module](https://www.financi.co.il/en/crm).

## A calculation model: what it is worth to you

The model below is based on the figures you enter, and is not a statistic about the market or other agencies. It is meant to illustrate the direction, so the conversation in the team is about your numbers:

*Interactive tool on the web page: Lead response-time model*

The result only means something if your input assumptions are real. Try a few scenarios: what happens if the conversion rate of quickly answered leads is ten points higher, and what if it is the same? Sometimes “almost no change” is the important result, because it shows the problem lies elsewhere. You can also run the model [as a standalone page](https://www.financi.co.il/en/tools/lead-response-time-model).

## A 20-minute weekly routine

1. **Open an exceptions report** — Leads that waited longer than the threshold you set, sorted by hour, source and owner.
2. **Ask “why?” not “who?”** — Look for a pattern: hour, day, channel, a stand-in who was not there. The aim is to change a rule, not blame a person.
3. **Change one rule** — Routing, fallback, a template or an on-duty hour. One change a week, so you know what worked.
4. **Check conversations with no owner** — On Friday, every open conversation has someone responsible. If not, assign it before the weekend.
5. **Record the decision** — What changed, from when, and what is expected to improve, so you can check it next week.

## What a useful exceptions report looks like

A good report is short and action-oriented. It does not show every lead, only the ones that exceeded the threshold, so the weekly meeting deals with problems rather than scrolling a table. The columns worth having:

| Column | Why it is there | Example use |
| --- | --- | --- |
| Lead source | Shows whether the problem is in a specific channel | Form leads are answered more slowly than WhatsApp leads |
| Arrival hour and day | Reveals time patterns | Every Thursday-afternoon lead waits until Sunday |
| Owner | Shows load and specialism | A lead that reached an employee on holiday |
| Time to assignment | Separates a routing problem from a response problem | The lead waited long before it had an owner |
| Time to first response | The main measure | Compare with the threshold you set |
| Current status | Shows whether the lead is still alive | A lead answered late and closed |

Set the threshold from your own data. In the first weeks set it so only a few leads exceed it, and tighten it gradually as the rules improve. That keeps the report a working tool and not a list of blame.

## The first WhatsApp message: a template ready in advance

A lead who left details in a form usually has not yet written to you on WhatsApp, so the first message to them comes from your side. According to Meta's developer documentation, a template is the only type of message that can be sent outside a customer service window, and only an approved (APPROVED) template can be sent. Review is automatic and can take up to 24 hours, and editing an existing template triggers a new review. So the template is written and approved before the first lead arrives, not while they wait.

- **One template per lead source,** so the message fits the context in which the lead left their details.
- **Correct classification.** The documentation requires every template to be classified as authentication, marketing or utility, and the category affects pricing. Do not mix marketing content into a service template.
- **A short template:** an acknowledgement, the owner's name, and what happens now.
- **Do not edit in the middle of a campaign,** so it does not go back into review.
- **One person approves new wording,** so a template does not stall between two people who each assume the other dealt with it.

Measure the template too: how many leads replied after it, and how fast. A template nobody answers is a sign to rewrite it, not a reason to change channel.

Add to the exceptions report any lead whose first message did not go out because there is no approved template, so this delay is visible in the weekly meeting and not discovered from the client.

## What the law says about marketing messages to a lead

Speed does not excuse the consent rules. Section 30A of the Communications (Telecommunications and Broadcasting) Law prohibits sending an advertisement by fax, automatic dialling, an electronic message (a message sent over the internet, which includes email) or SMS without the recipient's prior express consent in writing. The law makes an exception for an existing customer on three cumulative conditions: the recipient gave their details during a purchase or negotiation for one and the advertiser told them the details would be used to send advertisements, was given a chance to refuse and did not, and the advertisement concerns a similar type of product or service. Every advertisement must state the right to refuse and how to do it, and for a knowing breach the court may award damages that do not depend on proof of loss, up to ₪1,000 per advertisement. Acknowledging an enquiry the lead sent themselves is not the same as a campaign, but that distinction, and how it applies to WhatsApp, is a matter for legal advice and not an assumption for the sales team.

To stay out of trouble: separate operational messages (confirmation, scheduling, a document) from marketing messages, keep a record of when and how the lead agreed to receive messages, and add an opt-out to every marketing message.

## Common mistakes

- **Setting a target before measuring.** A copied number does not reflect your agency.
- **Counting an automatic confirmation as a response.** The client is waiting for a person, not an automatic reply.
- **Looking only at the average.** The average hides the cases that lose clients.
- **Routing without a fallback.** A rule without “what if they do not answer” moves the delay to one person.
- **Blaming people instead of changing a rule.** A recurring delay is usually a process problem.

## How to measure response time without fooling yourself

A single number for response time can hide more than it reveals. A few measurement rules prevent a picture that looks too good:

- **Define one start point:** the moment the lead is captured in the system, not the moment someone notices it.
- **Count the first human response:** an automatic message is not a response. It can come before it but not replace it.
- **Separate working hours from the rest:** a lead that arrived at night and was answered in the morning is not the same as a lead that waited an hour in the middle of a working day.
- **Look at the median and the tail:** the average is hidden by a few fast replies. The lead that waited longest is the one worth checking.
- **Split by source:** different sources behave differently, and an overall measure hides where the problem is.

> **Consistency before precision.** The measurement point matters less than consistency. Whoever changes the definition midway loses the ability to compare with last week.

## Frequently asked questions

### Does an automatic message count as a response?

Not for measurement. It can tell the client the enquiry was received, but a response is a human reply that moves the enquiry forward. Measure the time to the first human reply.

### How often should we review response time?

In a short weekly routine, and separately when something changes: a new lead source, a new employee or a change in working hours. A change without a check is a guess.

### What counts as a good response time?

We have no well-founded general number, so we will not publish one. Measure the current state, set a target that suits your team and hours, and check whether it improves.

### Does an automatic message count as the first response?

Not for measurement. It is good for confirming receipt, but the measure is the first human response.

### How do we handle leads that arrive at night?

An orderly confirmation saying when you will answer, and a morning queue that starts with the accumulated leads, so the first to respond is the right person.

### Does automatic routing suit a small agency?

In an agency of two or three people a written rule and an on-duty owner may be enough. Automatic routing matters when the team grows and leads arrive in several channels.

### How do we know whether speed improves conversion?

Compare leads that were answered quickly with those that waited, from the same source in the same period. The model shows the effect under your assumptions and does not replace checking real data.

> **Want to see lead routing and a shared inbox in one system?** Book a demo and we will show a lead arriving, getting an owner and being measured, including what Financi does not do. [Book a demo](https://www.financi.co.il/en#contact)

## Sources and further reading

- [Meta for Developers: WhatsApp message templates (documentation)](https://developers.facebook.com/documentation/business-messaging/whatsapp/templates/overview)
- [Communications (Telecommunications and Broadcasting) Law, section 30A: the text (Wikisource; check against the current text)](https://he.wikisource.org/wiki/%D7%97%D7%95%D7%A7_%D7%94%D7%AA%D7%A7%D7%A9%D7%95%D7%A8%D7%AA_(%D7%91%D7%96%D7%A7_%D7%95%D7%A9%D7%99%D7%93%D7%95%D7%A8%D7%99%D7%9D))
- [gov.il: Israel government services and information](https://www.gov.il)

---

*The information in this article is general only and is not financial, pension, legal or tax advice, nor a substitute for personal advice. Demos and calculators are illustrative models: your numbers will differ.*