An outbound funnel review for LinkedIn takes 20 minutes when you ask five questions in a fixed order: are the sender accounts connected, how many leads did each campaign reach, how many connection requests were accepted, how many conversations got a reply, and which replies still wait for an answer. With the GTM Bud connector, one person asks all five in one chat with Claude or ChatGPT, and the campaign results carry the time they were calculated.
This article is for a RevOps manager at a B2B company with 30 to 150 people who owns the weekly pipeline meeting and gets the outbound numbers as screenshots. We built the review order below while we wrote a weekly review routine for our own Claude setup, so the limits named here are the ones we met.
The scene: three reps, three screenshots, one meeting at 9:30
Marit is the RevOps manager at an 80-person company that sells scheduling software to warehouses. The pipeline meeting is on Monday at 9:30. At 9:10 she has three screenshots in a team channel. One is cropped. One is from Friday. The third shows a different campaign than the one she asked about.
The problem is not the reps. A screenshot has no definition and no timestamp. Marit cannot tell if “replies” means every reply or only the good ones, and she cannot add three pictures together.
The company has one GTM Bud account with the three LinkedIn accounts of the reps connected to it. Marit is signed in to that account from Claude. So at 8:45 she opens one chat. The meeting gets one table from one source.
Which numbers belong in an outbound funnel review?
An outbound funnel review needs five stages, read from the top down, because each stage can only be judged when the stage above it is healthy. Low replies mean nothing if no messages went out.
| Order | Stage | The number GTM Bud returns | What a weak number tells you |
|---|---|---|---|
| 1 | Sending | Connection status and pause state of each sender account | Nothing below this line is a message problem |
| 2 | Reach | Leads reached, with a count for each channel | The campaign is short of leads or of volume |
| 3 | Acceptance | Connection requests accepted, divided by requests sent | The audience or the sender profile is off |
| 4 | Replies | Leads that replied, divided by Conversations | The first message does not earn an answer |
| 5 | Replies that wait | The list of replies that need an answer | A person is the bottleneck, not the campaign |
Two definitions matter before the first Monday. Leads reached are leads that got at least one connection request, direct message, InMail or email from the campaign. Conversations are leads that got a direct message, InMail or email, or that replied. A connection request alone does not start a conversation. The reply rates use Conversations as the denominator, so they are not “replies divided by requests sent”.
The 20 minute agenda, minute by minute
Minutes 0 to 2: are the sender accounts connected?
Start here because a disconnected account explains every weak number below it. Marit types:
“List the connected sender accounts. Tell me which ones are not connected, which ones are paused, and which LinkedIn accounts have no Sales Navigator.”
The answer is a short list with three facts for each account. If one account shows a problem, the review of its campaigns is already done: fix the account first. Connecting an account is done on gtmbud.com, and the AI gives the link.
Minutes 2 to 8: results for each active campaign
“List my campaigns. For each active one, get the results. Give me one table: campaign, leads reached, acceptance rate with its two counts, conversations, total reply rate, interested reply rate, unlabeled replies, and the as-of time.”
The AI calls the results tool once for each campaign and builds the table. Each rate comes with the two counts behind it, so Marit can check the division. The numbers refresh when someone asks and can be up to about 15 minutes old. The as-of time gives the exact time.
Then the step that the screenshots never allowed:
“Here is the table from last Monday. Subtract it from the new one and show me the movement for each campaign.”
GTM Bud returns totals for the whole campaign, not for the last seven days. The movement of the week is the difference between two Mondays. Marit keeps each table in the meeting document, so the earlier table is one paste away.
Minutes 8 to 14: the replies that wait
This stage gets the most time. A lead who wrote on Thursday and has no answer on Monday is the most expensive line in the review.
“For each active campaign, list the replies that need an answer. Show the lead, the label, and the day of the last reply. Oldest first across all campaigns.”
The list shows each lead with a label of Interested, Warm or Not interested, or with no label yet. Marit does not answer the replies in this meeting slot. She counts them, notes the oldest, and takes the count to the meeting with one name next to it.
Minutes 14 to 17: the replies with no label
The results show a count of leads whose reply is not yet labeled. A high count can make the Interested number too low, because an interested lead with no label is not counted as interested.
“Open the unlabeled replies in the mid-market campaign one by one. Tell me the label you would give each and why. Do not set a label until I say so.”
A label is a write action. Marit reads the proposals, says which ones to apply, and the AI sets them.
Minutes 17 to 20: one change, written down
The last three minutes produce one sentence for the meeting. One change, one owner, one campaign. Not three.
What each stage tells you to change
Read the table from the top, and stop at the first weak stage. That stage is the one to fix this week.
- Sending is broken. Reconnect or unpause the account on gtmbud.com. Do not touch targeting or messages.
- Reach is flat between two Mondays. Check the campaign controls: the sending mode and lead finding. A campaign in the mode that sends only follow-ups reaches nobody new, by design.
- Acceptance is weak, reach is fine. The people are wrong or the sender profile does not interest them. A connection request from GTM Bud has no note, so the profile is all they see. Ask the AI to list the assigned leads and read twenty titles. If a third are outside the audience, the criteria need work. From the chat, targeting can be edited on a draft only. For a live campaign, change the targeting on gtmbud.com, or duplicate the campaign and give the copy new criteria.
- Acceptance is fine, replies are weak. People accept and then stay silent. The first message after the acceptance is the suspect. Read the saved sequence with the campaign sequence tool.
- Replies are fine, the waiting list is long. The campaign works and the team does not keep up. This is a staffing line for the meeting, not a campaign change.
The stop rule: fix the highest weak stage only. A change at two stages in one week makes the next review unreadable, because you cannot tell which change moved the number.
What this review cannot show you
Three limits, so that nobody promises them in the meeting.
- No number per rep in the chat. No tool returns which sender account a campaign or a reply uses. Marit names each campaign with the rep in it, so “per campaign” is “per rep” in practice. For a campaign with two or more LinkedIn senders, the campaign dashboard on gtmbud.com has a By sender table.
- No meetings and no pipeline value. The results stop at labeled replies. Meetings booked and opportunity amounts live in the calendar and the CRM. If Marit has a CRM connector in the same chat, her AI can read those there. GTM Bud does not.
- No InMail reply rate by itself. Reach is split by channel. The reply labels are not split between connection requests and InMails.
The table that goes into the meeting
Marit pastes one block into the meeting document: the results table with the as-of time, the count of waiting replies with the age of the oldest, and one line that starts with “This week we change”. The reps see the same table she sees. The argument about whose screenshot is right does not happen, because there is one source and it has a timestamp.
For the full list of what the chat can read and change, see the GTM Bud connector page. If your team has not yet decided who may approve what, start with the RevOps checklist for adding AI to outbound.
Frequently asked questions about a weekly outbound review
How long should a weekly outbound funnel review take?
Twenty minutes for three to five campaigns, if the order is fixed and the numbers come from one source. Most of the time goes to the replies that wait for an answer, because that is the one stage where a delay costs a real conversation.
Can I get last week only, and not the totals since launch?
Not from the tool directly. GTM Bud returns totals for the whole campaign. Paste the table from the earlier review into the chat, and your AI subtracts the old totals from the new ones to show the movement of the week.
Can the review show numbers for each rep?
In the chat, only if each rep has a separate campaign. No connector tool returns which sender account a campaign or a reply uses. Put the name of the rep in the campaign name, and the review per campaign becomes a review per rep.
Does the review change anything in my campaigns?
No, if you ask only the five questions in this article. Each one uses a tool that is marked read only. A change, such as a new label or a targeting edit on a draft, is a separate request that you make after the review.
Can the review run by itself each Monday?
GTM Bud does not start it for you. A person opens the chat and asks. Some AI apps have their own task timers, and whether yours can call a connector on a timer depends on that app, so test it before you rely on it.
Run the first outbound funnel review this Monday
An outbound funnel review works when the order never changes: sending, reach, acceptance, replies, replies that wait. Add the connector with the steps on GTM Bud for Claude, ask the five questions, and keep the table. The second Monday is the first one with a movement column. For the exact definition of each number before it goes on a slide, read the LinkedIn outbound metrics RevOps should report, and for the wider picture see AI LinkedIn outreach for RevOps on GTM Bud.