Campaign and Lead Desk Official

Every weekday morning it reads yesterday's spend and results for each live paid campaign, updates the campaign, and drafts a pause note when spend or cost per lead is off pace. It never pauses anything.

1. 01Campaign {{campaign}} on {{channel}}, budget {{budget}}, running {{start}} to {{end}}, spend so far {{spend}}, leads so far {{leads}}. Read the channel's platform and target cost per lead from the Channels table. If the channel is paid (TikTok or Bing), read yesterday's and the last three days' spend, clicks and leads from the ad platform. Compute the straight-line budget for the days elapsed and the three-day cost per lead. If the channel is email or organic, count the leads in the Leads table whose Source Campaign is this campaign instead. Output: The numbers you read, as plain text lines, or "no platform data" and why.
2. 02Update the campaign: Spend To Date, Leads, Cost Per Lead. Then apply the playbook's pace rules. If on pace, write one line: "On pace: <spend> of <straight-line>, CPL <x> against <target>". If off pace, write the pause note: what is off, by how much, and the one change you would make. Never change Status. Plain text. Output: The one-line pace note or the pause note, plain text.

Reads the draft in the email tool and checks every item on the playbook's QA checklist: links, personalisation fallbacks, unsubscribe, audience size, subject length, from name. Writes what passed and what failed. Never sends.

1. 01Send {{send}} for campaign {{campaign}} in {{email_tool}}: draft "{{draft_name}}", audience "{{audience}}" (planned size {{audience_size}}), planned subject "{{subject}}". Read the draft from the email tool. List its subject, preview text, from name, audience and size, every link, and every personalisation tag with whether it has a fallback. Fetch each link and note any that do not resolve or point to a staging domain. Output: The facts about the draft as plain text lines, with each link and its result.
2. 02Go through the playbook's QA checklist. Write one line per item: PASS or FAIL and why. If every item passes, set the send's Status to "QA passed"; if any fails, set it to "QA failed". Fill Subject and Audience Size from the draft if they were blank. End with one line: "Ready to send" or "Fix N items before sending". Never send or schedule. Plain text. Output: The checklist result, one line per item, plain text.

When a lead arrives it checks the CRM by email and domain, applies the routing rules to pick an owner, writes the contact to the CRM with its source, and says what it did. Never emails the lead.

1. 01Lead {{name}} <{{email}}> at {{company}}, from campaign "{{source_campaign}}" via {{form}}. Message: {{message}} Search the CRM by email, then by the company's domain. Note whether the contact exists, whether the company is a customer, and who owns either. Output: What the CRM holds for this person and company, plain text.
2. 02Apply the playbook's routing rules. Set In CRM Already, Owner and Status on the lead (Routed, Not a fit or Duplicate). For a new or existing contact, create or update the CRM contact with source campaign and form as properties and put the CRM link on the lead. Add one to the source campaign's Leads count. Write the routing result: the decision, the rule, the owner, the CRM link. Never email the lead. Plain text. Output: The routing result, plain text.

Loads each page, times it, checks the form is present and submits, reads the last seven days of visits and conversions from analytics, and flags slow pages, broken forms and conversion drops.

1. 01Page {{page}} at {{url}} for campaign {{campaign}}. Last check: {{visits}} visits, {{conversions}} conversions. Fetch the page, time it, look for the form and its submit button. Read the last 7 days of visits and conversions from analytics. Output: Load time, form present or not, visits, conversions, as plain text lines.
2. 02Update Visits 7d, Conversions 7d, Conversion Rate, Load Seconds and Last Checked. Set Status per the rules (Healthy, Slow, Form broken, Conversion drop). Write two lines: what you found and what to do about it. Plain text. Output: The two-line check, plain text.

Every Monday it reads position changes for your tracked queries and pages from Semrush, writes the movers into SEO Watch with a concrete suggestion for each, and drafts three content ideas.

1. 01Week of {{week}}. Read the tracked queries and pages from Semrush with position, weekly change and clicks. List the five biggest gainers and the five biggest losers. Output: The movers with their numbers, plain text.
2. 02Create one SEO Watch row per mover (Item, Kind, Position, Change, Clicks 7d, Suggestion, Week, Status New) with a specific suggestion each. Then write three content ideas the numbers support, one line each. Plain text. Output: The rows created and the three ideas, plain text.

Counts spend, leads and cost per lead by channel for last week, names the best and worst campaign, lists three decisions, writes the scorecard row and emails it to the marketing lead.

1. 01Week of {{week}}. From the tables: spend and leads by channel for last week, cost per lead per channel and per campaign, sends that went out with opens and clicks, landing pages not Healthy, leads not yet routed. Output: The numbers as plain text lines.
2. 02Write the scorecard: the numbers as short lines, "Best:" and "Worst:" campaign with cost per lead, "Watch:" the pages and sends with issues, "Decisions this week:" three one-sentence decisions. Create the Weekly Scorecard row. Send it with Turtle Notify to the marketing lead. Under 250 words, plain text. Output: The scorecard as sent, plain text.