Direct dial phone numbers are the single highest-leverage variable in outbound calling, and they are also the field where B2B data vendors are weakest. Coverage claims are inflated, sourcing methods are rarely disclosed, and the legal obligations attached to the number change depending on whether it rings a desk or a pocket.
The performance gap is not subtle. Teams calling switchboard numbers typically connect on 3-5% of dials. Teams working verified direct lines and mobiles report 8-12% and 18-22% respectively. Same script, same rep, same hour of the day. Only the number changes.
This guide explains how direct dial phone numbers are actually produced, why coverage collapsed after 2020, how to model the real cost per conversation, where mobiles beat direct lines, and which rules apply before you press dial.
Direct dial, switchboard and mobile: three different assets
The three number types are often lumped together in a single “phone” column, which is how teams end up dialling a receptionist for a quarter. They behave nothing alike.
| Number type | What it reaches | Typical connect rate | Gatekeeper risk | Regulatory weight |
|---|---|---|---|---|
| Switchboard / HQ | Main reception | 3-5% | High | Lowest |
| Direct dial (desk) | An extension assigned to one person | 8-12% | Low | Moderate |
| Verified mobile | The individual, wherever they are | 18-22% | None | Highest |
That last column is the trade-off most vendor content ignores. A mobile number reaches the buyer directly, but it also pulls you into stricter consent and dialler rules in the United States. More reach means more responsibility, not less.
How direct dial phone numbers are actually sourced
Vendors describe this as “our proprietary data network”, which is a sentence designed not to answer the question. In practice, direct dial phone numbers come from six identifiable supply routes, and each has a different reliability profile.
| Source | How it works | Reliability | Main weakness |
|---|---|---|---|
| Human phone verification | Researchers call and confirm the person answers | Highest | Expensive, so coverage is narrower |
| Opt-in and first-party submissions | Event registrations, gated content, profile updates | High | Skews to marketing-engaged personas |
| Public filings and registries | Corporate records, licensing bodies, regulators | High | Mostly senior and regulated roles |
| Contributory networks | Users share address books in exchange for credits | Medium | Consent lineage is often thin |
| Extension inference | Known switchboard plus known extension pattern | Medium-low | Fails after any phone system migration |
| Scraped signatures and directories | Harvested from web pages and documents | Low | Stale, and legally the most exposed |
Ask any vendor for the percentage split across these six routes. Most will not have the number to hand, which is itself informative. The ones who do will usually be the ones investing in verification. Our B2B data accuracy testing framework includes this question in its RFP list for exactly that reason.
Why desk-line coverage collapsed, and what replaced it
Hybrid work quietly destroyed a large slice of the direct-dial universe. Companies retired desk handsets, moved to softphones, and stopped publishing extension ranges. A number that once rang a specific desk now rings an empty office or a decommissioned VoIP endpoint.
As a result, two things happened at once. Desk-line accuracy fell, and the strategic value of mobile numbers rose sharply. Vendors that built their reputation on extension inference have struggled since, while those investing in human phone verification and mobile coverage have pulled ahead.
For buyers, the practical implication is simple: stop evaluating “phone coverage” as one number. Insist on separate fill rates for direct dial and mobile, per region. A vendor at 70% blended phone coverage might be at 65% desk lines and 12% mobiles, which is a very different product from the one you thought you were buying.
The real economics: cost per conversation, not cost per record
Procurement teams compare data vendors on price per record, which is the wrong denominator. The correct one is cost per conversation, because rep labour dwarfs data cost. Here is the model on 100 dials, using a fully loaded SDR cost of roughly $38 per hour and 2.5 hours per 100 dials.
| Scenario | Data cost / 100 | Labour cost / 100 | Connects | Cost per conversation |
|---|---|---|---|---|
| Switchboard only | $0 | $95 | 4 | $23.75 |
| Direct dials at $0.15/record | $15 | $95 | 10 | $11.00 |
| Verified mobiles at $0.40/record | $40 | $95 | 20 | $6.75 |
The conclusion holds across almost any set of realistic inputs: the most expensive data is nearly always the cheapest outcome, because you are buying back rep hours. Free switchboard numbers are the most expensive option on this table by a factor of three.
Run this model with your own comp figures before your next renewal. It reframes the negotiation from “your CPM is high” to “show me your connect rate on my ICP”, which is a conversation good vendors welcome.
When mobiles win, and when direct dials still win
Mobile is not universally better. The right default depends on the persona and the sector, and this is the nuance missing from most published guidance.
- Mobiles win for field-based roles, construction, logistics, hospitality and multi-site retail, where nobody sits at a desk.
- Mobiles win for founders and C-level buyers at companies under 200 people, where the mobile is the business line.
- Direct dials win in regulated environments such as banking, government, healthcare and defence, where personal mobiles are often prohibited or firmly off-limits.
- Direct dials win for mid-level operational roles in large enterprises, who still work from assigned extensions.
- Switchboards still have a job: confirming a person is in role before you spend a better number on them.
Build your dial strategy per segment rather than per team. Pair it with timing signals, since reaching the right person at the wrong moment converts poorly. We covered that in B2B buying signals and in speed to lead.
The legal layer before you press dial
Calling rules differ by jurisdiction, by number type and by the technology you dial with. Get this wrong and the penalties are per call, which scales badly. What follows is orientation, not legal advice.
United States: TSR, DNC and TCPA
The FTC’s Telemarketing Sales Rule contains a business-to-business exemption at section 310.6(b)(7), which removes most TSR obligations from genuine B2B calls, with narrow exceptions such as nondurable office and cleaning supplies. The National Do Not Call Registry likewise does not cover business numbers called for a business purpose. The full text sits in the FTC’s guide to complying with the Telemarketing Sales Rule.
That exemption is narrower than it sounds. You must still transmit accurate caller ID, and you must still maintain and honour an internal do-not-call list. More importantly, the TCPA governs calls and texts to mobile numbers made with an automatic telephone dialling system or a prerecorded message, and B2B status does not switch that off. Since April 2025, revocation of consent must also be honoured promptly and across channels.
- Manual dialling of a business direct line: lowest risk.
- Autodialled calls or texts to a personal mobile: highest risk. Take advice first.
- State-level rules: several states diverge from the federal B2B exemption. Check your calling territories individually.
United Kingdom and EU
In the UK, corporate subscribers can register with the Corporate Telephone Preference Service, and screening against CTPS and TPS is mandatory before cold calling. Across the EU, the number itself is personal data, so you need a documented lawful basis under Article 6 of the GDPR, plus the notice and objection handling that goes with it. The ICO’s guidance on legitimate interests sets out the three-part test you should document.
We break the whole picture down, country by country, in our guide to B2B data compliance.
How to evaluate a phone data vendor
- Ask for direct dial and mobile fill rates as separate numbers, split by country.
- Ask what share of numbers were human phone-verified in the last 90 days.
- Ask which DNC, TPS and CTPS registries are screened, and how often.
- Sample-dial 100 records before signing and log wrong-person, dead-line and connect outcomes.
- Ask for the correction SLA when you report a bad number, and get it into the contract.
- Ask how numbers are classified. A vendor that cannot tell a mobile from a desk line cannot help you stay compliant.
Point six is the quiet one. If mobile and desk numbers arrive in the same undifferentiated field, your dialler configuration cannot apply different rules to them, and your compliance posture is guesswork.
Dial hygiene that protects your connect rate
- Vary the time of day. Calling the same slot repeatedly reaches the same voicemail repeatedly.
- Cap attempts. Six to eight touches over two to three weeks, then rest the record.
- Use local presence honestly. Accurate caller ID is a legal requirement, not a growth hack.
- Log every outcome. Wrong-number outcomes are your best signal of vendor decay.
- Re-verify quarterly. Direct lines decay at roughly 25-30% a year, as covered in our CRM data hygiene guide.
How ZenBee sources direct dial phone numbers
ZenBee separates desk lines from mobiles in the data model, so your dialler and your compliance rules can treat them differently. Numbers are continuously re-verified across a network of 700M+ profiles across 35M+ companies, and phone coverage is reported per region rather than as a single flattering global average.
Because dialling, sequencing and signals sit in one platform, a wrong-number outcome updates the record immediately and feeds the next refresh. If you are still stitching a phone vendor to a dialler to a CRM, see how the ZenBee sales platform collapses that stack, or read our primer on choosing a B2B prospecting tool.
Frequently asked questions about direct dial phone numbers
What is the difference between a direct dial and a mobile number?
A direct dial is a desk extension assigned to one person, so it bypasses reception but only reaches them at their desk. A mobile reaches the individual anywhere, which raises connect rates and also raises the regulatory bar, particularly under the TCPA in the United States.
What is a good connect rate for direct dial phone numbers?
Expect 8-12% on verified desk lines and 18-22% on verified mobiles. Anything at or below 5% usually means you are dialling switchboards, stale extensions, or the wrong persona entirely.
Is it legal to call direct dial phone numbers without consent?
In the United States, genuine B2B calls fall under a TSR exemption, though internal do-not-call lists and accurate caller ID still apply, and mobiles dialled automatically remain subject to the TCPA. In the UK and EU you need a documented lawful basis and must screen against the relevant preference services. Take local advice before you scale.
Why does my phone data get worse over time?
Desk lines decay at roughly 25-30% a year because people change roles and companies retire handsets. Mobiles decay far more slowly, at around 10-15%, because the number travels with the person rather than the job.
Are free sources of direct dials worth using?
Signatures, filings and company sites can yield accurate numbers, and they are excellent for spot research. At team scale, however, the labour cost of manual sourcing exceeds the data cost of a good vendor several times over, as the cost-per-conversation model above shows.