
LinkedIn tells you who to call. It almost never tells you on which number. To get from a profile to a phone that rings, you take the profile URL, run it through a contact data source that returns a mobile number, and test that number before you build a session on it. Here is the full path, ranked by what actually connects.
A LinkedIn profile has a "Contact info" section. Members can add a phone number there, and they decide who sees it. In practice very few do, and those who do usually show it to their connections only.
That is by design. LinkedIn is a network for identity and context, not a phone book. What it gives you is still valuable for a call:
So treat LinkedIn as the starting point of the search, not as the place where the number lives.
We rank these by a single criterion: how often the number you get ends in a conversation with the right person.
This is the main method. You give a data provider the LinkedIn URL, or the name and company, and it returns the professional contact details it holds for that person, ideally a direct mobile.
Coverage varies a lot by country, seniority and company size, and every provider is stronger in some markets than others. In our experience, French and European mobile coverage is where providers differ the most. Test on a sample of your own accounts before you commit.
No single source has everyone. A waterfall queries a first source, then a second for the contacts the first one missed, then a third.
It raises coverage, and it raises the need for testing, because you now mix numbers of different quality in one list. Keep the source of each number in the CRM so you can compare pickup rates later.
The cheapest number is the one you already have. Old deals, email signatures, past call logs, inbound forms. Reps use only about 20% of the signals already in their CRM, and phone numbers are part of that.
Check the contact and their colleagues. A number that answered two years ago is still the best lead you have on that account.
If you are connected, open "Contact info". It takes five seconds and occasionally works.
You can also ask. A short message after a real exchange, "happy to explain in five minutes, what is the best number?", gets a number that the person expects you to use. It does not scale, and it has the best pickup rate of any method.
The main line is public and easy to find. It is also the slowest path: a gatekeeper, a transfer, often a voicemail. Use it for accounts that matter when you have nothing else, not as the base of a session.
In our benchmark of 1,000,000 B2B cold calls in France, the average connect rate is 11.3%, and it ranges from 11% to 20% depending on persona and time slot.
The data behind those calls matters as much as the slot. A direct mobile rings in the prospect's pocket. A switchboard number rings at a desk where someone's job is to filter you. In our experience, the share of direct mobiles in a list is what separates a good session from an empty one.
So do not measure a data source by how many numbers it returns. Measure it by how many of those numbers produce a conversation. We go deeper on this in our article on the real cold call connect rate.
A number found is not a number that works. People change jobs, companies change operators, and some numbers in any database were never right.
Three checks, from cheapest to most reliable.
Log wrong numbers as wrong. A "wrong person" outcome that stays in the CRM as a valid mobile will be called again by the next rep.
We are not lawyers and this is not legal advice. Two points every team should take seriously.
B2B prospecting has to respect GDPR. A professional's mobile number is personal data. In practice that means working with data sources that can explain where their data comes from, telling the people you call who you are and how you got their details if they ask, and honouring every request to stop. Rules on phone prospecting also differ from one European country to another, so check with your legal counsel for each market you call.
Scraping is a risk. LinkedIn's User Agreement prohibits automated collection of member data. Tools that harvest profiles at scale expose your account to restriction and your company to questions you do not want. Looking at a profile and enriching a contact through a provider that holds its own data is a different thing from extracting LinkedIn's.
The practical rule: if you cannot explain to a prospect, in one sentence, how you got their number, do not use that source.
Our view is simple: a number is only worth its pickup rate. Sonik is an AI sales dialer, so the place where we judge contact data is the calling session, not the spreadsheet.

The parallel dialer gives reps about five times more conversations, and it makes data quality visible fast: the live statuses show which numbers ring, which connect and which are dead. A weak list shows in the first half hour of a session.
The LinkedIn context is not lost either. Job changes and social intent are among the signals Sonik reads, next to website visitors, closed-lost deals and dormant leads, so the reason you looked at the profile becomes the opener of the call.
LinkedIn gives you the person and the reason. The number comes from enrichment, your own CRM or a direct ask, and it has to be tested before it is trusted. Favour direct mobiles, keep the source of every number, and stay away from scraping. This week, take 100 contacts from your current list, tag each number with its source, call them, and compare the pickup rate per source.