indranilbanerjee/digital-marketing-pro / skills/context-engine
context-engine
VerifiedLoad shared marketing context: brand profile, benchmarks, compliance rules, specs. "load my brand context"
execution-workflows.md
View on GitHub# Execution Workflows — Standard Operating Procedures
All marketing execution follows standardized workflows to ensure consistency, quality, and compliance. Each workflow is a numbered step-by-step procedure that agents follow during execution. Skipping steps is not permitted — if a step does not apply, log it as "N/A" with a reason.
---
## 1. Blog Publishing Workflow
Use for: blog posts, articles, guides, listicles, how-to content, thought leadership.
### Content Preparation
1. **Format check** — Verify content is in the target CMS format (HTML for WordPress, rich text for Webflow). Strip unsupported elements.
2. **Word count validation** — Confirm word count meets the content brief. Minimum thresholds by type:
- Standard blog post: 800-1,500 words
- Pillar page: 2,000-4,000 words
- News/update post: 400-800 words
3. **Readability score** — Run through readability analysis. Target Flesch-Kincaid grade level based on audience:
- B2C general: Grade 6-8 (Flesch score 60-70)
- B2B professional: Grade 10-12 (Flesch score 40-55)
- Technical/developer: Grade 12-14 (Flesch score 30-45)
4. **Brand voice check** — Score content against brand voice profile. Minimum score: 70/100.
### SEO Optimization
5. **Primary keyword placement** — Verify primary keyword appears in:
- Page title (within first 60 characters)
- H1 heading (exactly once)
- First 100 words of body text
- At least one H2 subheading
- URL slug
6. **Meta description** — Write or verify meta description: 150-160 characters, includes primary keyword, contains a clear value proposition or CTA.
7. **Alt text** — Every image has descriptive alt text. Include primary keyword in at least one image alt (naturally, not forced).
8. **Internal links** — Minimum 3 internal links to relevant existing content. At least 1 link in the first 300 words. Anchor text is descriptive (not "click here").
9. **External links** — Include 1-3 authoritative external sources where relevant. Set to `target="_blank" rel="noopener"`.
10. **Schema markup** — Apply appropriate JSON-LD schema (`BlogPosting`, `Article`, or `HowTo`).
### Publishing
11. **Category and tag assignment** — Assign to exactly 1 primary category and 2-5 relevant tags. Use existing taxonomy; create new terms only when no match exists.
12. **Featured image** — Upload featured image meeting platform specs (WordPress: 1200x628 px minimum). Verify it displays correctly in social share preview.
13. **Schedule or publish** — If scheduling, confirm publish date/time aligns with content calendar. If publishing immediately, confirm with user.
14. **Verify live URL** — After publishing, confirm the live URL responds with HTTP 200. Check that the page renders correctly (no broken images, no layout issues).
15. **Log execution** — Record in campaign tracker: URL, publish date, primary keyword, word count, readability score, SEO score.
16. **Monitor first 24h** — Check pageviews, bounce rate, and average time on page after 24 hours. Flag if pageviews are below 50% of the brand's blog average.
---
## 2. Email Campaign Workflow
Use for: marketing emails, newsletters, drip sequences, promotional blasts, event invitations.
### List Selection
1. **Segment identification** — Select the target segment based on campaign brief. Document segment name, size, and selection criteria.
2. **List size verification** — Confirm recipient count. Flag if count deviates more than 20% from expected size.
3. **Consent verification** — Verify all recipients have valid marketing consent per applicable regulation:
- GDPR regions: Opt-in consent on file
- CAN-SPAM regions: No opt-out on file
- CASL regions: Express or valid implied consent (check expiry)
4. **Suppression check** — Cross-reference against suppression list (unsubscribes, bounces, complaints). Remove any matches.
### Template Build
5. **Subject line** — Write subject line: 30-50 characters for mobile safety. Score against email scoring rubric (minimum 70/100).
6. **Preview text** — Write preheader: 40-90 characters. Must complement (not repeat) the subject line.
7. **Body structure** — Build email body: clear hierarchy (headline, body, CTA), single-column for mobile-first, inline CSS for compatibility.
8. **CTA placement** — Primary CTA button above the fold. Repeat CTA at email bottom. Button minimum 44x44 px tap target.
9. **Personalization mapping** — Map merge tags to data fields. Test that all merge tags resolve correctly. Set fallback values for empty fields.
10. **Unsubscribe link** — Confirm functional unsubscribe mechanism is present and visible. One-click unsubscribe header included.
### Quality Assurance
11. **Spam score check** — Run content through spam analysis. Target SpamAssassin score below 5.0. Flag trigger words (free, guarantee, act now, limited time).
12. **Test send** — Send to internal seed list (minimum 3 addresses across Gmail, Outlook, Apple Mail).
13. **Rendering review** — Verify rendering on desktop and mobile. Check dark mode compatibility. Confirm images load and alt text displays when images are blocked.
### Send
14. **Approval gate** — Obtain explicit approval from authorized user before sending.
15. **Schedule or send** — Set send time via the ladder: the ESP's per-recipient STO if available, else send-time-optimizer.py with `--history` (the list's own send log), else its dated population baseline. Respect quiet hours (no sends 9 PM-8 AM recipient local time).
16. **Monitor delivery** — Track within first 2 hours:
- Deliverability: Target >95%
- Bounce rate: Flag if >3%
- Unsubscribe rate: Flag if >0.5%
- Spam complaints: Flag if >0.1%
17. **Performance log** — Record: send date, list size, subject line, open rate, click rate, conversion rate, revenue attributed.
---
## 3. Ad Campaign Workflow
Use for: Google Ads, Meta Ads, LinkedIn Ads, TikTok Ads, Pinterest Ads, Amazon Ads.
### May 2026 platform updates (read before configuring)
- **Google Performance Max (PMax) 2026 updates** — brand exclusion lists are now first-class (use them to keep PMax from spending on competitor brand terms or off-strategy categories); per-network placement reporting now exposed (use this to verify YouTube vs Display vs Search distribution matches the campaign objective); first-party audience exclusions supported; **up to 15 videos per asset group** (support.google.com/google-ads/answer/14528532, read 2026-10-28); **PMax experiments** for testing creative or audience changes against a control. Combined effect: PMax is now reasonable to run if you instrument it; the old "black box" criticism is mostly addressed by the exclusion + reporting features.
- **Meta Advantage+ shopping 2026** — in-app checkout, AI product overlays (price/reviews on product hover), retailer integrations standard. Creative should be designed at the product-tile scale, not full-frame. Catalog quality dominates performance — invest in clean product feed (high-res images, accurate stock status, full attributes) before tuning bids. **API note (checked 2026-10-27):** standalone ASC campaigns can no longer be created via the Marketing API (all versions, since 19 May 2026), and v26.0 (introduced 29 Jul 2026) blocks edits to every legacy ASC/AAC campaign and pauses the ASC campaigns that use `existing_customer_budget_percentage` — configure shopping automation through the unified Advantage+ setup on standard campaigns.
- **Meta Advantage+ Leads — globally available (May 2026)** — Meta's automated lead-generation campaign type rolled out globally in May 2026 (was country-limited in 2025). Single campaign type that auto-tests creative, audience expansion, and placement. **Use when:** lead-gen is the objective AND the brand has installed lead-quality feedback (Conversions API tied to CRM) — without quality feedback Advantage+ Leads optimises for cheap leads, not qualified leads. **Don't use when:** the brand is in a regulated Special Ad Category (housing, credit, employment) — automated audience expansion fights against the manual targeting constraints those categories require.
- **Threads ads** — Threads ads run globally. Formats, from the Threads API changelog (developers.facebook.com/documentation/threads/changelog, read 2026-10-27): image ads, image carousels (2025-10-06), Advantage+ catalog ads (2025-10-28), app ads (2026-02-17), and promoting existing Facebook or Instagram posts from page-backed accounts (2026-04-21); video ads since 2025-08-15 (REPORTED). Inventory is cheap-to-moderate (still ramping). Best for: brand awareness with a younger / news-adjacent audience, journalism / B2B thought-leadership amplification, and creator collaborations. Pair with Instagram in the same campaign for cross-app delivery. Our opinion, not a platform rule: keep Threads to a small share of Meta budget until your account-level data shows comparable CPA.
- **Meta brand-safety controls (May 2026)** — Advertiser-level inventory filters were tightened in Q2 2026 (reported). Three tiers: Expanded (most reach, lowest brand-safety filter), Moderate (default), and Limited (highest filter, least reach). Default to Moderate; move to Limited only for regulated industries (finance, healthcare, B2G) or after a documented brand-safety incident — the reach cost is real. The Brand Suitability dashboard now exposes which Reels topics your ads served adjacent to in the last 30 days.
- **LinkedIn Ads** — LinkedIn has published no ranking statement with a number. Practice: links in the post body may reduce organic reach, so test per account; in-platform formats (lead gen forms, document ads, conversation ads) keep the user on LinkedIn. LinkedIn's marketing blog (linkedin.com/business, 2025-09-24) announced Qualified Leads Optimization, Conversion Value Optimization and Book-an-Appointment. Lead gen forms remain the highest-converting LinkedIn ad format.
- **TikTok Ads (post-USDS Jan 2026)** — US-served ads now run through USDS LLC infrastructure; ad-account migration was automatic for existing US advertisers. AI-generated creative requires AI disclosure label.
- **Retail media (Amazon, Walmart, Instacart)** — combined retail-media ad spend now ~$60B+/year (US 2026). Worth a campaign track for any DTC or CPG brand selling on these platforms.
### Campaign Structure
1. **Campaign hierarchy** — Set up: Campaign (objective, budget, schedule) → Ad Group/Ad Set (targeting, bid) → Ad (creative, copy, CTA, destination URL).
2. **Naming convention** — Apply standard naming: `{Brand}_{Objective}_{Audience}_{Platform}_{Date}`. Example: `Acme_Leads_Retargeting_Meta_2026-02`.
### Audience Setup
3. **Targeting definition** — Configure audience: demographics, interests, behaviors, geo-locations. Document all targeting parameters.
4. **Exclusions** — Set exclusion audiences: existing customers (if prospecting), recent converters, internal traffic, competitors' employees (if possible).
5. **Lookalike/similar audiences** — If using, specify source audience (minimum 1,000 contacts for Meta, 300 for LinkedIn) and similarity percentage (1-3% for precision, 5-10% for reach).
### Creative
6. **Creative upload** — Prepare assets per platform specs (see `platform-specs.md` and `platform-publishing-specs.md`). Minimum 3 creative variations for A/B testing.
7. **Ad copy** — Write copy per platform character limits. Score against ad creative rubric (minimum 70/100 to launch).
8. **Landing page** — Verify destination URL loads in under 3 seconds, matches ad messaging (message match), and has working conversion tracking.
### Budget and Bidding
9. **Bid strategy selection** — Choose strategy aligned with campaign objective:
- Awareness: Maximize reach / Target CPM
- Traffic: Maximize clicks / Target CPC
- Conversions: Maximize conversions / Target CPA / Target ROAS
10. **Daily budget** — Set daily budget within brand's `budget_range`. Never exceed the brand's maximum daily budget without explicit re-confirmation from user.
11. **Budget pacing** — Configure even daily distribution unless campaign brief specifies otherwise (e.g., dayparting, accelerated delivery).
### Compliance and Launch
12. **Compliance review** — Check against platform ad policies (see `compliance-rules.md` Section 4). Verify required disclaimers for regulated industries. Confirm Special Ad Category designation if applicable (housing, credit, employment).
13. **Conversion tracking** — Verify pixel/tag fires correctly on conversion page. Test with a click-through before launch.
14. **Launch** — Activate campaign. Confirm status shows "Active" or "Running" (not "Limited" or "Error").
### Optimization
15. **48-hour check** — Review: impressions, CTR, CPC. Pause underperforming creatives (CTR below 50% of ad group average). Confirm budget is pacing correctly.
16. **7-day optimization** — Review: CPA, ROAS, conversion rate. Adjust bids, reallocate budget from underperforming ad groups. Add negative keywords (search campaigns).
17. **14-day review** — Full performance review: creative fatigue check (frequency >3.0), audience saturation, budget efficiency. Refresh creatives if CTR has declined >20% from launch.
18. **Performance log** — Record: impressions, clicks, CTR, CPC, conversions, CPA, ROAS, spend. Compare against industry benchmarks from `industry-profiles.md`.
---
## 4. Social Scheduling Workflow
Use for: organic social media posts across all platforms.
1. **Content format** — Prepare content per platform specifications (see `platform-specs.md`). Respect character limits, image sizes, and video duration limits per platform.
2. **Platform-specific formatting** — Adapt content for each platform's culture:
- LinkedIn: Professional tone, industry insights, paragraph breaks
- Instagram: Visual-first, caption with line breaks, emojis appropriate to brand voice
- Twitter/X: Concise, conversational, thread format for long content
- TikTok: Native video, trending audio consideration, casual tone
3. **Hashtag research** — Select 3-5 hashtags per post. Mix: 1-2 high-reach (100K+ posts), 2-3 niche/industry-specific (1K-100K posts). Platform exceptions: Twitter/X (1-2 max), LinkedIn (3-5), Instagram (5-10).
4. **Optimal posting time** — Use `posting-time-analyzer.py` with `--history` (the brand's engagement export) when available, else its dated population baseline; the brand's established measured times outrank both. Cross-reference with `platform-specs.md` general recommendations.
5. **Visual assets** — Verify all images/videos meet platform dimension and file size requirements. Check text-safe zones for Stories/Reels content.
6. **Schedule** — Queue post in scheduling tool or platform native scheduler. Confirm scheduled time is correct (timezone awareness).
7. **Engagement monitoring** — Check engagement at 2 hours and 24 hours after posting. Respond to comments within brand guidelines.
8. **Performance report** — At 24 hours and 7 days, log: impressions, reach, engagement rate, saves, shares, link clicks (if applicable). Compare to brand's rolling 30-day average.
---
## 5. CRM Operations Workflow
Use for: contact imports, lead creation, data updates, list management, pipeline operations.
1. **Data validation** — Check all records for required fields. Validate formats:
- Email: RFC 5322 compliant, domain has MX record
- Phone: E.164 format (international), strip non-numeric characters
- URL: Must include protocol (https://)
- Country/State: ISO 3166 codes
2. **Deduplication check** — Run dedup using matching hierarchy:
- Primary: Email address (exact match)
- Secondary: Phone number (normalized)
- Tertiary: Company name + contact name (fuzzy match, >85% similarity)
3. **Field mapping** — Map source fields to CRM native fields. Document any unmapped fields. Create custom fields only with explicit approval.
4. **Pre-import snapshot** — Take a backup snapshot of affected records before any bulk operation. Store snapshot path in execution log.
5. **Import/create** — Execute import. For bulk operations (>100 records): use batch API endpoints, implement rate limiting per CRM vendor specs.
6. **Verification** — Confirm: record count matches expected count, no data truncation, all required fields populated. Spot-check 5 random records for accuracy.
7. **Campaign linking** — Associate imported contacts with the relevant campaign/list in CRM.
8. **Sync log entry** — Record: operation type, record count, source, date, operator, any errors or warnings.
---
## 6. Report Delivery Workflow
Use for: weekly pulse reports, monthly reviews, QBRs, ad-hoc performance reports.
1. **Data pull** — Query all connected analytics sources (Google Analytics, Search Console, ad platforms, email platforms, CRM) via MCP servers. Normalize date ranges to the reporting period.
2. **Aggregation** — Combine data from multiple sources. Normalize: currency (convert to brand's primary currency), date formats (ISO 8601), metric definitions (e.g., "conversions" definition consistent across platforms).
3. **KPI calculation** — For each KPI, calculate:
- Current period value
- vs. target (% attainment)
- vs. previous period (% change, direction arrow)
- vs. industry benchmark (above/below, percentile if available)
4. **Anomaly detection** — Flag any metric that changed >25% period-over-period. Flag any KPI below 80% of target. Flag any metric that deviates >2 standard deviations from the 90-day mean.
5. **Format selection** — Choose delivery format per report type:
- Weekly pulse: Slack blocks or short email
- Monthly review: Google Slides or HTML email with charts
- QBR: Google Slides presentation deck
- Ad-hoc: Google Sheets or direct message
6. **Brand/agency voice** — Apply appropriate voice:
- Internal reports: Can use shorthand, reference multiple clients (agency mode)
- Client-facing reports: Professional, third-person, brand-appropriate language
7. **Send** — Deliver report via configured channel. Include a 2-3 sentence executive summary at the top.
8. **Confirm receipt** — Log delivery timestamp. For email delivery, verify no bounce. For Slack, verify message posted successfully.
---
## 7. SMS/WhatsApp Workflow
Use for: promotional SMS, transactional SMS with marketing elements, WhatsApp marketing messages.
### Compliance (Mandatory First Step)
1. **Consent verification** — Verify opt-in consent for every recipient:
- US (TCPA): Express written consent for marketing SMS. Prior express consent for informational. Honour a revocation made by any reasonable means within 10 business days; from 25 Nov 2026 a revocation may be scoped to the category of informational messages it answered, and you may designate one exclusive way to revoke (91 FR 67214). A STOP to marketing ends marketing. See `compliance-rules.md` section 1.2.
- EU (GDPR): Explicit opt-in consent.
- Canada (CASL): Express consent with documented proof.
- All: Consent must be specific to SMS/WhatsApp channel (email consent does not transfer).
2. **Quiet hours** — Enforce sending windows: no messages between 9:00 PM and 8:00 AM in recipient's local timezone. South Korea: no messages 9 PM-8 AM (law). US: varies by state, 8 AM-9 PM is safest.
3. **Opt-out mechanism** — Every message must include opt-out instructions:
- SMS: "Reply STOP to unsubscribe" (exact or equivalent)
- WhatsApp: Unsubscribe link or reply keyword
### Message Preparation
4. **Message formatting** — Respect character limits:
- SMS (GSM-7 encoding): 160 characters per segment. Unicode: 70 characters per segment.
- WhatsApp: 1,024 characters. Supports rich media (images up to 5 MB, video up to 16 MB, documents up to 100 MB).
5. **Sender ID verification** — Confirm sending number/ID is registered and verified:
- SMS: Toll-free number, short code, or 10DLC (US). Alphanumeric sender ID (international, where supported).
- WhatsApp: Verified WhatsApp Business API number with approved display name.
6. **Template approval (WhatsApp)** — WhatsApp marketing messages require pre-approved templates. Submit template for Meta approval (24-48h turnaround). Session messages (within 24h of user message) do not require templates.
### Send and Monitor
7. **Test send** — Send to internal test numbers first. Verify delivery and rendering.
8. **Send** — Execute send. Monitor delivery rate in real-time.
9. **Delivery tracking** — Track: delivery rate (target >95%), response rate, opt-out rate (flag if >1% per send). Log undeliverable numbers for list cleanup.
10. **Cost tracking** — Record per-message cost. SMS costs vary by country and carrier. **WhatsApp Business Platform switched to per-message pricing on 1 July 2025** (was per-conversation). India marketing-template messages are roughly USD 0.0118 each (cheapest globally; Germany ~11×). Click-to-WhatsApp ads and Facebook Page CTAs open a 72-hour free service window; service messages inside the 24-hour customer-care window remain free. Alert if campaign cost exceeds budget by >10%.
---
## 8. Memory Operations Workflow
Use for: storing marketing insights, campaign learnings, content to brand memory, knowledge base updates.
1. **Content extraction** — Strip formatting artifacts. Extract key points, metrics, and actionable insights. Preserve source attribution (campaign name, date, channel).
2. **Metadata tagging** — Apply structured metadata:
- `content_type`: insight | campaign_data | performance_snapshot | content_piece | voice_sample
- `tags`: Array of relevant topic tags (e.g., ["email", "subject-lines", "open-rate"])
- `source`: Origin campaign or analysis
- `date`: ISO 8601 timestamp
- `confidence`: high | medium | low (for insights)
3. **Deduplication** — Generate content hash. Check against existing entries. If duplicate detected, update the existing entry's metadata (timestamp, confidence) rather than creating a new entry.
4. **Storage** — Write to appropriate location under `~/.claude-marketing/brands/{slug}/`:
- Insights: `insights.json` (rolling buffer, max 200 entries, oldest evicted)
- Campaigns: `campaigns/{id}.json`
- Performance: `performance/{campaign}-{date}.json`
- Content: `content-library/`
5. **Index update** — Update relevant index files (`_index.json` for campaigns). Ensure the new entry is searchable by tags and date.
6. **Verification** — Read back the stored entry. Confirm content matches what was written. Confirm search/lookup returns the entry.
7. **Sync state** — Update `_active-brand.json` last-modified timestamp. Log the memory operation in the execution log.
---
## Pre-Flight Checklist
A quick-reference table of mandatory checks per execution type. All checks must pass before execution proceeds.
| Check | Blog | Email | Ads | Social | CRM | Report | SMS/WA | Memory |
|---|---|---|---|---|---|---|---|---|
| Brand context loaded | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Compliance rules applied | Yes | Yes | Yes | Yes | Yes | No | Yes | No |
| Content scored (rubric) | Yes | Yes | Yes | Yes | No | No | No | No |
| Platform specs verified | Yes | Yes | Yes | Yes | No | No | Yes | No |
| Consent/opt-in verified | No | Yes | No | No | No | No | Yes | No |
| Budget within range | No | No | Yes | No | No | No | Yes | No |
| Test send/preview | No | Yes | No | No | No | No | Yes | No |
| Approval gate | Medium | Medium | High | Medium | High | Low | High | Low |
| Backup/snapshot taken | No | No | No | No | Yes | No | No | No |
| Tracking verified | Yes | Yes | Yes | No | No | No | Yes | No |
| Performance log entry | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |