Run outbound with one founder LinkedIn account

Most outbound advice assumes a team. SDR org, sales engineer, a deck designer, three connected accounts, an inbox manager. That advice is mostly useless if you are a founder doing outbound from one LinkedIn account because you have not hired yet, or because you are deliberately staying lean.

The good news: a single founder account, used well, can produce a steady pipeline without putting the account at risk. The bad news: most of the standard playbook scales down badly. This is the one that does.

Step 1. Use your real account, not a fake one

The temptation is to spin up a separate “outreach” LinkedIn account so the real one stays clean. Resist. Fake accounts get flagged faster, accept rates run lower because there is no profile to validate against, and you lose the credibility that a real founder profile carries.

What matters is that the account you use looks like a founder doing founder things: complete profile, recent posts, a few comments on real conversations each week, real connections. If your profile is half-finished, fix it before you turn outbound on. The accept rate delta between a complete profile and a sparse one is roughly fifteen points.

Step 2. Set the weekly cap at forty, not eighty

Funkel’s default is eighty invites per sender per week. That cap is tuned for a sender that has been warmed for a month and is being run by someone whose only job is outbound. A founder account that is also doing customer support, hiring conversations, fundraising messages, and personal posts is signaling LinkedIn in a different shape than a pure outbound account.

Lower the cap to forty. That works out to roughly eight invites a day across five business days, which leaves the account headroom for the other things you actually use it for. Reply rates compensate; on a tight signal, forty invites a week from a real founder account out-converts eighty from a generic SDR account.

Step 3. Schedule around your real founder activity

Two scheduling rules.

Match the activity window to your real LinkedIn habit. If you post in the morning and reply to comments in the afternoon, set the campaign window to 9:00-13:00 in your timezone. Outbound that fires at 3am local time is the loudest signal of automation; firing during your real activity window blends in.

Post once a week, comment three times a week. Non-outbound activity is what tells LinkedIn you are a real person, not a bot wearing a profile. The ratio matters: an account sending forty invites and posting nothing looks suspect; the same account posting once and commenting a few times looks normal.

Funkel paces sends with a 30-180s random delay between actions and respects the per-campaign business hours. You do not need to add throttling on top.

Step 4. Pick one signal that suits founder voice

Founder accounts have one structural advantage: the buyer knows you are the founder, not a junior SDR. Use it. The signals that work hardest with founder voice:

• your_profile_url: people who looked at your profile. Volume is low (5 to 15 a week for most founders) but reply rate is the highest of any signal because they already came to you.
• competitor_pain: people publicly complaining about your competitor. Founder voice opens this conversation more credibly than an SDR voice (“I built X because [the same problem]”) so use that framing.
• recent_job_changes: use as a baseline, not the primary. New role buyers expect a founder note to feel personal, so the AI prompt should lean conversational, not structured.

Pick one for the first month. Add a second only after the first signal has produced a stable two weeks of replies.

Step 5. Use AI mode, but train the tone explicitly

Manual templates from a founder account read as templates-from-a-founder-account, which is worse than templates-from-an-SDR. Buyers hold founders to a higher voice bar.

AI mode handles this if you give it concrete voice instructions. In the agent’s tone settings:

• Paste three actual messages you have written (not marketing copy, real DMs to peers).
• Ban specific phrases you would never say (“synergize”, “circle back”, “touch base”).
• Specify length: one sentence for invites, two to four for follow-ups, no exceptions.

Then turn review mode on for the first week. Approve twenty messages by hand. After the twentieth, you will know whether the model has your voice; turn review mode off and let it run.

Step 6. Add a second sender when, not before

The Growth plan includes two senders by default. The temptation is to use both immediately. Do not. The second sender adds capacity at the cost of attribution: you can no longer tell which messages are landing because of who the sender is versus what the message says.

Add the second sender when one of three things is true.

• The first sender is consistently capped at forty invites a week and the analytics show a queue of unsent candidates building up.
• You hired or partnered with someone whose LinkedIn presence credibly matches your category (a first SDR, a co-founder, a known operator who agreed to lend their account).
• You are running two distinct ICPs that justify two different voices. One sender per ICP keeps each voice clean.

If none of the three is true, the second sender slot is better left unused. Funkel does not bill differently for running one sender versus two; you are paying for capacity that may not need to be filled yet.

What to skip when you are solo

• A/B testing message variants. You do not have the volume to make the test mean anything; pick one and commit.
• Multi-step sequences longer than three messages. Each additional step at low volume produces noise, not insight.
• Running outbound on a signal whose volume is below ten candidates per week. The fixed cost of attention is higher than the upside.
• Adding agents for products you do not yet have a clear ICP for. Two agents on Growth is more than enough; the third costs attention you do not have.

For pricing details on running this on Growth, see our pricing.