Part 6 — Measurement: Knowing What Works

Part 6 of 7 in the Your AI VP of Marketing series. Previous: Part 5 — The Pipeline | Next: Part 7 — Scale: Automation, Sales, and the Pattern

Measure Before You Automate

Your VP has been shipping work. The pipeline from Part 5 generates briefs, checks voice, plans campaigns, repurposes content, and schedules distribution. Content is landing. Channels are warming up. If this were a human hire, you would now be sitting down for the first quarterly review — the conversation where you stop asking "are you producing?" and start asking "is any of it working?"

That is what this Part does. You are going to hold your VP accountable to numbers.

The temptation is to skip to automation — set the machine on a timer and focus on the rest of the business. That is Part 7. Do not skip to it yet. Automating without measurement is like doubling your ad spend without checking which ads convert. More content, faster, across more channels, with no idea whether any of it is working. You will not know which channel brings leads, which content format gets shared, or whether your email list is growing or bleeding subscribers. Speed without direction is expensive noise.

This Part gives you direction. You will define the metrics that matter, build a Skill that analyzes them, store your baseline in Memory, and set a cadence that tells you when to celebrate, when to adjust, and when to kill something that is not working. Once you know what success looks like, automating becomes a force multiplier instead of a gamble.

Define Your KPIs

KPIs — key performance indicators — are the numbers that connect marketing activity to business outcomes. The phrase sounds corporate. The concept is simple: pick the handful of numbers that tell you the truth, and watch them over time. Pick the three to five that match the goals you set in your CLAUDE.md back in Part 1.

Website sessions by source. How many people visit, and where do they come from? Four sources matter: organic search, social, direct, and referral. If you are posting on three social platforms but 80% of traffic comes from organic search, that is a signal about where your effort should go.

Email open rates and click-through rates. Open rate tells you whether your subject lines work. Click-through tells you whether your content compels action. Industry averages hover around 20-25% open rate and 2-3% click-through for small businesses, but your own trend matters more than the benchmark. A 32% open rate dropping to 28% over three months is more useful than knowing the industry average.

Social media engagement. Likes, comments, shares, and follower growth — broken out by platform. Follower count alone means nothing. A thousand followers with 1% engagement produce less than three hundred at 5%. Track engagement rate (interactions divided by followers) and watch the trend month over month.

Leads by channel. Where do your actual leads come from? Not visitors — people who filled out a form, booked a call, subscribed to a trial, or emailed about pricing. This separates marketing activity from marketing results. If LinkedIn generates three times more leads than Instagram despite half the posting frequency, that changes your entire distribution strategy.

Revenue by channel. If you can trace a sale back to a channel, this is the metric that ends all debates. A six-month sales cycle with five touchpoints gets messy. But if you sell online, use discount codes, or ask "how did you hear about us?" on intake forms, you have enough to make it useful.

Match the metrics to your CLAUDE.md goals. Growing your email list? Email metrics and lead sources are primary. Driving traffic from search? Organic sessions and keyword rankings matter most. Building community on social media? Engagement rates are your north star.

Pick three to five. Write them down.

Build a Metrics Analysis Skill

You built Skills in Part 3 — the Content Brief Generator, the Brand Voice Checker. Same pattern here, but this Skill does something different: it reads performance data and tells you what the numbers mean.

The easy path: /skill-creator. From Part 3 onward, /skill-creator is the canonical route for every new Skill. Open a new conversation in your Cowork project, type /skill-creator, paste the prompt below, and answer the interview questions. On approval, the builder writes .claude/skills/metrics-analyzer/SKILL.md for you. The manual paste-block is preserved in the "Under the hood" callout further down if you want to compare what /skill-creator generates to the hand-built version.

Paste this into /skill-creator:

Build a skill called "Marketing Metrics Analyzer" that:
- Accepts marketing performance data (CSV, pasted table, or text summary)
- Compares current metrics to baseline values stored in Memory
- Identifies trends (improving, declining, flat) for each metric
- Highlights the top 3 wins and top 3 concerns
- Recommends specific actions based on the data
- Outputs a structured report with sections for each channel
- Flags any metric that has dropped 20% or more from baseline
- Notes any metric that has improved 20% or more from baseline

/skill-creator will ask you follow-up questions. Here are the answers you will likely need:

What inputs does the skill need? Marketing performance data in any format — CSV, a pasted table, or a plain text summary. The skill should handle whatever the user gives it.

Should it reference anything from Memory? Yes. Baseline metrics stored in Memory. The skill compares current data to the baseline to identify trends.

What format should the output take? A structured report with these sections: Executive Summary (3-4 sentences), Channel Performance (one subsection per channel), Top 3 Wins, Top 3 Concerns, Recommended Actions, and a Data Quality note flagging any missing or suspicious numbers.

How should it handle missing data? Note the gap and analyze what it has. Do not refuse to produce a report because one metric is missing.

Review the generated skill, adjust anything that does not match your needs, and save it.

Analyze the marketing performance data provided and produce a structured
performance report.

INPUTS
- Marketing data: CSV, table, or text summary (provided by user)
- Baseline metrics: retrieved from Memory entry "kpi-baseline"

ANALYSIS PROCESS
1. Parse the provided data and identify all available metrics
2. Retrieve baseline values from Memory
3. Calculate period-over-period changes for each metric
4. Compare current values to baseline
5. Identify trends: improving (3+ months of growth), declining (3+ months
   of decline), or flat
6. Flag any metric that has moved 20%+ from baseline in either direction

OUTPUT FORMAT

## Marketing Performance Report — [Date Range]

### Executive Summary
3-4 sentences covering overall trajectory, biggest win, biggest concern,
and one recommended priority action.

### Channel Performance
One subsection per channel (Website, Email, Social, Leads, Revenue).
For each channel:
- Current metrics vs. baseline
- Trend direction and magnitude
- One-sentence assessment

### Top 3 Wins
The three metrics showing the strongest positive movement. For each:
what improved, by how much, and what likely caused it.

### Top 3 Concerns
The three metrics showing the most worrying trajectory. For each:
what declined or stalled, by how much, and what might be causing it.

### Recommended Actions
3-5 specific, prioritized actions based on the data. Each action should
name the channel, the metric it targets, and the expected impact.
"Post more on social media" is not specific enough. "Increase LinkedIn
posting from 2x/week to 4x/week to test whether engagement rate
scales with frequency" is.

### Data Quality Notes
Flag any missing metrics, suspicious numbers (e.g., a 500% spike that
might be a tracking error), or gaps in the data that limit the analysis.

RULES
- Compare to baseline from Memory. If no baseline exists, note this and
  analyze the data on its own merits.
- Be direct about what is working and what is not. Do not soften bad news.
- Every recommended action must tie to a specific metric and channel.
- If a metric has dropped 20%+ from baseline, flag it prominently.
- If a metric has improved 20%+ from baseline, call it out as a win.

Baseline Your Current State

A trend requires at least two data points. Before the metrics analyzer can tell you whether things are improving, it needs to know where you started. That starting point is your baseline, and you are going to store it in Memory — the same persistent context store you first worked with in Part 1 — so your VP can reference it in every future analysis.

Gather your current numbers. If you have real data from Google Analytics, your email platform, and your social media dashboards, export the last three months. If you are following along with the tutorial and want to test the skill first, use this sample data for a fictional business:

month,website_sessions_organic,website_sessions_social,website_sessions_direct,email_subscribers,email_open_rate,email_click_rate,instagram_followers,instagram_engagement_rate,linkedin_followers,linkedin_engagement_rate,leads_organic,leads_social,leads_referral,revenue_marketing_attributed
2026-01,1240,380,520,1850,34.2,4.1,2100,3.2,890,4.8,12,4,8,4200
2026-02,1310,420,490,1920,32.8,3.9,2180,3.0,920,5.1,14,5,7,4800
2026-03,1450,510,530,2010,35.1,4.5,2290,3.4,960,4.6,18,7,9,5600

A small business doing reasonably well. Organic traffic growing. Social sessions climbing. Email open rates healthy but click-through inconsistent. Instagram followers up, engagement mediocre. LinkedIn smaller but higher engagement. Leads trending upward across all channels. Revenue growing month over month. Not perfect numbers — real numbers never are. The mix of strong and weak channels is what makes this useful for testing.

Paste the CSV into your Cowork project with this prompt:

Store this as our marketing baseline in Memory entry "kpi-baseline". These
are our Q1 2026 metrics. Use the March values as the primary baseline for
future comparisons, and note the January-March trend for each metric.

[paste your CSV here]

After storing, confirm what you saved and highlight any metrics that
already show a clear trend direction.

Your VP will parse the data, store the baseline values in Memory, and give you a summary of what it sees. The March numbers become your reference point. The three-month window gives trend context.

From here on, every time you run the metrics analyzer skill, your VP pulls these baseline values from Memory and compares new data against them. You do not re-paste the baseline. It persists. This is the feedback loop: the pipeline produces, the analyzer reads the output, and results flow back into Memory — where the next run will read them as context. Memory is not just storage. It is the place where your VP learns what worked.

The Reporting Framework

Data without rhythm becomes noise. You check your numbers once, get a snapshot, and forget about it until something breaks. A reporting framework turns measurement from an event into a habit.

Three cadences cover everything you need.

Weekly: The Pulse Check

Once a week, spend five minutes on the basics. You are looking for anomalies, not trends. Did traffic drop off a cliff on Wednesday? Did an email blast produce zero clicks? Is there a social post that got ten times the normal engagement?

Prompt your VP:

Here is this week's data: [paste quick numbers or a screenshot summary]

Give me a 3-sentence pulse check. Anything unusual? Anything I should
act on this week?

The weekly check is not an analysis. It is a smoke detector. If nothing unusual happened, the answer is "steady week, no action needed" and you move on.

Monthly: The Channel Review

Once a month, run the full metrics analyzer skill with your latest data. Compare current performance to baseline, look at trends, and decide whether to adjust.

Run the metrics-analyzer skill with this data:

[paste your monthly CSV or data summary]

Compare against our stored baseline. What improved? What declined?
What do you recommend we change?

The monthly report is your primary decision-making tool. It tells you where to double down, experiment, or cut losses. If LinkedIn leads are up 40% from baseline while Instagram engagement is flat, that is a resource allocation decision staring you in the face.

Quarterly: The Strategy Review

Every three months, step back from channel-level data and ask bigger questions. Are the goals from your CLAUDE.md being met? Has the market shifted? Should you add a channel, drop one, or change your content mix? This is the real quarterly review — where your VP earns its keep as a strategist, not just an analyst.

Review our last three monthly reports and our original marketing goals
from CLAUDE.md. Are we on track? What should change for next quarter?

Produce a quarterly strategy brief with:
- Progress against each goal (on track / behind / ahead)
- One thing to stop doing
- One thing to start doing
- One thing to keep doing

The quarterly review connects data back to business goals and recommends changes to the plan, not just the tactics. It also writes new context back into Memory — what worked, what did not, what the next quarter should prioritize. That is how the VP gets smarter over time: each review becomes the baseline for the next one.

What triggers action between cadences? Three scenarios should prompt an immediate review, regardless of the schedule:

• Any metric drops 20% or more from baseline in a single period. Something broke and you need to find out what.
• A channel consistently underperforms for three consecutive months. Flat is fine for a month. Flat for a quarter means the channel is not working.
• A channel suddenly outperforms expectations. A metric jumping 30% above baseline is not just good news — it is a signal to investigate what caused it and do more of that.

Off-Ramp 3: What You Have Built

What you have built: A complete content marketing pipeline with measurement — you know what is working, what is not, and where to focus. This is a professional-grade marketing system.

What is ahead: Part 7 automates recurring work, extends the pattern into sales, and teaches you to expand the system yourself. But your pipeline is already measurably producing results.

This is a real stopping point, and a good one. Look at what you built across six Parts. Your VP knows your business, your audience, and your voice. It follows brand standards and process rules on every task. The pipeline from Part 5 generates briefs, plans campaigns, repurposes content, and distributes it on a schedule. And now it measures the results.

Most small businesses do not have this. They post on social media because they feel like they should, send emails when they remember, and check analytics once a quarter when someone asks "how is marketing going?" You have a system, data, and a VP that tells you what the data means and what to do about it.

If you stop here, you are running a more measured marketing operation than companies with ten times your budget and a full-time team that never looks at the numbers. The pipeline produces. The measurement tells you whether the production matters.

What just changed

Before this exercise, your Cowork Project shipped content but had no repeatable way to measure outcomes. You walked /skill-creator through the metrics-analyzer specification and approved the save to the skills directory. Cowork wrote .claude/skills/metrics-analyzer/SKILL.md and stored your opening numbers in the Memory entry kpi-baseline. Future reporting runs load the Skill, compare new numbers against that baseline, and flag real movement. The feedback loop is closed: pipeline output becomes Memory input, and the next review starts where this one left off.

What is Next

Your pipeline works. You can measure it. The next step is to stop initiating every task by hand.

Right now you run the pipeline when you open Cowork. You paste data when you remember. You trigger the distribution planner when you have time. Every piece works, but it all depends on you remembering to start it.

Part 7 changes that. You will learn Scheduled Tasks — a Cowork feature that triggers recurring work on a cadence you define. The metrics analyzer you just built is a prime candidate: on the first of every month, your VP pulls the latest data, compares it to the baseline, writes the report, and surfaces the top wins and concerns before you open the project. The same Part extends the pattern into sales and closes out the series.

Part 6 of 7 in the Your AI VP of Marketing series. Previous: Part 5 — The Pipeline | Next: Part 7 — Scale: Automation, Sales, and the Pattern