How Accurate Is Call Location Routing?
How we work out where a caller is, how accurate each method really is, and the three ways to set up a campaign when your destinations can only take calls from certain areas. Written for licensed verticals like insurance, legal, and home services.
Call Tracking is in beta and turned on per account. If you don't see Calls in your left sidebar (or it shows a "Soon" tag), email support@leaddistro.ai and we'll switch it on for you.
If your destinations can only take calls from certain places, this page is the one to read first. An insurance agent is licensed in specific states. A law firm is admitted in specific states. A contractor only drives so far. Sending them a call they can't legally or practically take wastes the call and can create a compliance problem. This page covers how we work out where a caller is, how accurate each method really is, and the three ways to build the campaign. For the step-by-step of setting territories see Routing to Local Agents by Territory, and for the ZIP capture methods see Collecting ZIP Codes on Calls.
Where a caller's location comes from
Every inbound call gets one location, worked out once when the call arrives and then used everywhere: the ring tree, the RTB auction, the IVR, and the call record. We build it from these sources, and the ones higher in the list win.
| Source | How exact | What it costs you |
|---|---|---|
| Sent ahead of the call by your traffic partner | Exact. It is whatever they collected. | Your partner has to send it. Nothing for the caller. |
| Typed by the caller on the keypad (Collect ZIP Code) | Exact for callers who complete the prompt. | A few seconds of call time. Some callers hang up. |
| The tracking number's region, when set to always use it | Exact for the market the number runs in. | One number per market, and a media buy split to match. |
| The caller's phone number (automatic) | Roughly 2 out of 3. See below. | Nothing. It is always there. |
The automatic guess is the weak one, and it is the default. A phone number tells you where the number was issued, not where the person is. People move, keep their number, and port it between carriers. On mobile and internet numbers it is wrong often enough to matter. Every call tracking platform has this same limit, so treat the automatic location as a starting point, not the answer.
When a caller types their ZIP, we also work out their state from it automatically. So a state based rule ("licensed in Texas") works off a typed ZIP, and you do not have to ask for the state separately.
Where accuracy actually costs you money
Before changing anything, look at your own coverage map. If several destinations cover the same area, a wrong guess is harmless, because any of them can take the call and the ring tree just picks the next one. A wrong guess only costs you where one destination covers an area on its own. Those are the areas worth protecting.
- Covered by several destinations: a wrong location is harmless. Priority order decides, and everyone who could take the call is eligible anyway.
- Covered by exactly one destination: a wrong location sends the call to somebody who cannot take it, or drops it into your fallback. This is the money.
List the areas only one destination covers. If you can only fix accuracy for part of your traffic, fix it for those areas first.
The three ways to set up the campaign
These are the real choices. They differ mostly in who does the work: option 2 asks the caller, option 3 asks your media buyer.
| 1. One campaign, location targeting | 2. Add a ZIP prompt | 3. A number per market | |
|---|---|---|---|
| How it works | All destinations in one ring tree, each with its own area. We route on the automatic location. | Same as 1, plus the caller keys in their ZIP before the call routes. | A separate ad campaign and tracking number per market, each number locked to its region. |
| Accuracy | Roughly 2 in 3. | Exact for callers who complete the prompt, automatic guess for the rest. | Exact for the market. |
| Caller has to do anything? | No. | Yes, about 5 keypresses. | No. |
| Who sets it up | You, once. | You, once. | Your media buyer, ongoing. |
| Cost | None. | None. | A number per market, and one budget split several ways. |
| Time to go live | Minutes. | Minutes. | Days. New ad campaigns have to start pacing. |
| Calls lost to the setup | None. | Some callers hang up at the prompt. | None. |
| Areas covered by several destinations | Handled automatically. | Handled automatically. | Awkward. You have to assign the area to one campaign, or run it in several and bid against yourself. |
| When a destination is capped, busy, or closed | Falls through the ring tree. | Falls through the ring tree. | Weaker. The number points at one market, so you still need the others in the ring tree as backup. |
| Cost reporting | One pooled cost. | One pooled cost. | Ad spend per market, so cost per call per market. |
| Adding a destination later | Add it and set its area. | Add it and set its area. | New ad campaign and new number each time. |
Start with option 2. It is free, live in minutes, needs nobody outside your team, and keeps the ring tree fallback that makes a ring tree worth having. Move to option 3 for the markets where a single destination covers an area on its own and the volume justifies its own ad campaign. They stack: the number pins the market, the prompt confirms the caller.
Turning on the ZIP prompt
On the campaign's Call Flow tab, open the Collect ZIP Code step and switch it on. You can write your own prompt, or keep the default ("Please enter your 5 digit ZIP code"). The ZIP the caller enters routes the call, fills in their state, and is saved on the call for reporting. It is fail-open: a caller who says nothing still routes on whatever location we already had. Full detail in Collecting ZIP Codes on Calls.
Locking a tracking number to a market
On the campaign's Tracking Numbers tab, click a number's Region cell and set the state, and a ZIP if you want a tighter center. Then pick how firmly it applies.
- Fallback (default): use this region only when we can't tell where the caller is. Safe to set on any number.
- Always use this region: treat every call to this number as that market, whatever the caller ID says. Use this when the number runs in one market only.
Requiring a known location
Some calls arrive with no location at all. Each destination has a Require a known location switch inside its Location Targeting settings that decides what happens then. On, a call with no location skips that destination and moves to the next one in the ring tree. Off, it rings anyway. New destinations you add have it on, because a destination you drew a territory for is usually one that cannot take an out-of-area call. Destinations you set up before this switch existed keep ringing on those calls until you turn it on.
Leave at least one destination with this switch off. It is what catches the calls nobody else can take. If every destination requires a known location, a call we can't place reaches nobody and goes to your fallback message or voicemail.
When a call does skip a destination this way, the call's Timeline says "Caller location unknown" rather than "Location filter not matched", so you can tell a missing location apart from a real mismatch.
FAQ
Frequently Asked Questions
How accurate is the automatic caller location?
Should I build one campaign or one per destination?
Does the ZIP prompt lose me calls?
Do I need to collect the state as well as the ZIP?
What happens to a call we can't place at all?
Can I see how many of my callers are on mobile?
Related Articles
If you have any questions, send us an email at support@leaddistro.ai