I never trust one reply rate

A founder shows me their outreach dashboard on a call and starts with one reply rate for the whole campaign.

That number is not enough.

One number for every sequence, channel, and mailbox tells me how the campaign looks on average. But average is not something I can act on.

So I break it down.

Start with the sequence

Sequence-level performance gives me the top picture.

Contacts, replies, positive replies, and opportunities.

This tells me whether the campaign is moving in the right direction before I go deeper.

If the sequence itself looks weak, I know there is a bigger issue. If the sequence looks fine, I start checking what is driving the result.

Then check channels and mailboxes

Email and LinkedIn need to be checked separately.

LinkedIn can be starting conversations while email drags the number down. Other times email is doing the work and LinkedIn brings in nothing.

If you keep both channels grouped together, you miss that.

After channels, I check mailbox performance. One weak mailbox can quietly pull the sequence average down while everything above it still looks fine.

Check tests only if you are testing properly

If you are running tests, change one thing at a time.

One message. One step. One angle.

Change two things at once and you will not know what moved the number.

You can now pull this through Claude

We have Forge Stack MCP and CLI for this.

You can ask Claude to pull your sequence, channel, and mailbox numbers without clicking through multiple screens.

New marketplace deals added

We added new deals inside the Salesforge Software Discount Marketplace.