Multithreading by phone means calling three or more people in the same account within the same week, each with an opener built from the previous conversation, so that the deal never depends on one person picking up or staying in their job. It is a prospecting method, not an account-management one: the map is built before the first meeting, not after. Here is why single-thread outbound fails, who to call first, and the one-week sequence we use.

Why one contact per account is a bet you will lose

Gartner puts the typical B2B buying group at six to ten people. Most SDR sequences target one. The math is unkind: if that one person connects 11% of the time, the average across our benchmark of 1,000,000 B2B cold calls in France, the account has an 89% chance of staying silent on any given day, and a good chance of going dark for good when that person changes jobs, goes on leave or simply decides you are not their problem.

Single-threaded deals also stall later. The champion loves the demo, then disappears into an internal process you never mapped. In our experience the deals that close fastest are the ones where the second and third conversation happened before the first meeting, not after it.

Multithreading fixes both. More people called means more chances to connect this week, and more conversations before the meeting means the meeting is with the right people.

Who to call first: the pain owner, not the title

The usual advice is to climb the ladder or to start at the top. We start with the person who feels the problem every day and can name the others.

• The user owns the pain and knows the real process. Easiest conversation, most honest answers, and the source of the names you need next. The call's purpose is facts and names, not a meeting.
• The manager owns the number the pain affects. Purpose: confirm the cost of the problem in their words.
• The budget owner decides. Purpose: the meeting, opened on what the two previous people said.
• The blocker (IT, legal, security, procurement) is called last, before they are needed, with a question rather than a pitch.

Seniority is not the problem people think it is. Sales and leadership personas connect above 15% in our data, higher than the average. Executives pick up. They just do not give a stranger 20 seconds for a pitch. They give 20 seconds for "I spoke with Camille on your team yesterday about the onboarding backlog".

The one-week sequence

Everything in the same week. A conversation from three weeks ago is not a thread, it is a memory.

1. Day 1, the user. Open on the trigger (a hire, a job change, a visit to your pricing page). Ask one "how" question: discovery questions starting with "how" lift conversion by 190% in our data. End by asking who else lives with the problem. Write down the names.
2. Day 2, the manager. Open with the user's own words, not your product. "Your team described X yesterday, how do you see it from your side?" This is the trigger-based opener, and it books 7-10% of the time against 2-3% for a pitch-first open.
3. Day 4, the budget owner. Open on both conversations. Ask for a meeting with all three. A group meeting booked from three conversations rarely gets cancelled.
4. Day 5, the map. Log the roles, the quotes and the open questions in the CRM. Schedule the blocker call for the week of the meeting.

Time each call by persona. Founders pick up most around 10am, marketing around 4pm, HR around 1pm. Across personas 4pm is the best slot at 15.8%, 8am the worst at 9.8%. Three calls in one account can mean three different slots in two days.

What to say when the second person has never heard of you

They will not have. That is the point. The previous conversation is the credibility, not your logo.

The opener is a reference, not a name-drop: "I was talking with someone on the operations side here about X. Before I go further, how does that land on your side?" If they ask who, say who. If the first person asked you not to, do not call the second one yet.

Then stop talking. 65% of reps open by explaining what they sell. The multithreaded opener works precisely because it does not.

Letting AI pick the next person to dial

Mapping a committee by hand for every account is the reason reps do not multithread. It takes 20 minutes per account and the list has 50 accounts.

This is a job for AI, and a narrow one. Given a trigger on an account, the model should propose the three roles most likely to be involved, find the people holding those roles, get their mobile numbers through an enrichment waterfall, and write a one-line brief for each. The rep decides the order and makes the calls.

The second job is the hand-off between calls: after the user conversation, the AI summarises it and drafts the opener for the manager from the transcript. That is what makes day 2 possible without a spreadsheet.

How Sonik approaches multithreading

Sonik's account mapping starts from a signal: a visit, a job change, a closed-lost anniversary. The AI proposes the committee for that account, enriches the contacts and adds them to the rep's dialing session, each in the slot where that persona picks up.

After each call, the AI call summary becomes the brief for the next person in the account, so the rep opens on the previous conversation without rewriting notes. Managers see threads per account, not just dials per rep.

In our experience accounts worked with three contacts in the same week book meetings at a visibly higher rate than single-contact accounts, and those meetings hold.

Conclusion

One contact is a hope. Three contacts in one week is a process. Start with the person who owns the pain, open every following call on the previous conversation, and let AI do the mapping so the rep does the talking. This week, pick five target accounts and call three people in each before Friday. Count the meetings against your usual week. The difference is what single-threading has been costing you.