The Problem
A growth marketer at a 20-person SaaS startup sends 4-6 email campaigns per week across onboarding sequences, product updates, and promotional blasts. Campaign data is exported as CSV files from their email platform, but deep analysis never happens. The routine: glance at the open rate in the dashboard, decide it looks "okay" or "bad," and move on to the next send.
Six months of campaign data sits in a folder, and nobody has answered the basic questions: Which subject line patterns actually perform best? What send times drive the most clicks? Is there a drop-off point in the onboarding sequence where new users stop engaging? Decisions about what to send next are based on gut feeling, and the gut has no way to distinguish signal from noise across 142 campaigns.
The email platform dashboard shows metrics per campaign, but it cannot show patterns across campaigns. It cannot tell you that question-based subject lines consistently outperform declarative ones, or that Tuesday sends beat Friday sends by a wide margin. Those insights only emerge from cross-campaign analysis -- the kind nobody has time to do manually.
The Solution
Using the data-analysis, excel-processor, and report-generator skills, the agent merges 6 months of campaign CSVs, identifies the patterns behind top-performing emails, pinpoints timing and subject line strategies that drive engagement, and surfaces problems (like onboarding drop-offs) that are invisible from the email dashboard alone.
Step-by-Step Walkthrough
Step 1: Import and Normalize Campaign Data
I have 6 months of email campaign data exported as CSV files in the campaigns/ folder. Each file has: campaign name, send date, subject line, recipient count, opens, clicks, unsubscribes, and conversions. Analyze all of them together and tell me what's working and what's not.
The agent reads all 6 CSV files, merges them, and cleans up inconsistencies -- date format mismatches between monthly exports, a few duplicate campaigns from re-sends, and 12 campaigns with missing conversion tracking. The combined dataset:
- 142 campaigns from January through June 2026
- 847,000 total sends
- 24.3% average open rate
- 3.1% average click rate
- 0.8% average conversion rate
Those averages hide a wide range. The best campaign hit a 38.4% open rate. The worst: 11.2%. Understanding what separates them is the whole point of this analysis.
Step 2: Decode Subject Line Performance
Subject lines are the single biggest lever for open rates, and the patterns across 142 campaigns are remarkably clear:
Top-performing patterns:
| Pattern | Avg Open Rate | vs. 24.3% Baseline |
|---|---|---|
| Questions in subject line | 31.2% | +28% |
| Personalized with first name | 28.7% | +18% |
| Numbers/lists ("5 ways to...") | 27.9% | +15% |
Worst-performing patterns:
| Pattern | Avg Open Rate | vs. 24.3% Baseline |
|---|---|---|
| ALL CAPS words | 18.4% | -24% |
| "Newsletter" in subject | 19.1% | -21% |
| Over 60 characters | 20.2% | -17% |
The single best-performing subject line across all 142 campaigns: "Did you see this, {first_name}?" at 38.4% open rate. It combines two winning patterns -- personalization and a question -- while staying well under 60 characters.
The "Newsletter" finding is particularly useful. The monthly newsletter format has the lowest open rates of any campaign type, yet it consumes the most production time. It is a candidate for retirement or radical reformatting.
Step 3: Analyze Send Time and Day Performance
Break down performance by day of week and time of day. When should I be sending emails?
Best days by click rate:
| Day | Click Rate | Relative Performance |
|---|---|---|
| Tuesday | 4.1% | Best day by far |
| Thursday | 3.8% | Strong second |
| Wednesday | 3.2% | Solid mid-week |
Best send times by open rate:
| Time Window | Open Rate |
|---|---|
| 9:00-10:00 AM | 28.7% |
| 1:00-2:00 PM | 26.1% |
| 7:00-8:00 AM | 25.4% |
Avoid: Friday after 3 PM (17.2% open rate) and weekends (15.8% open rate). The data confirms what you might expect for a B2B SaaS audience -- people engage with work-related email during work hours, especially at the start of the workday and right after lunch. But now there are specific numbers instead of assumptions, and they reveal that the current Friday afternoon send habit is actively underperforming.
Step 4: Find the Onboarding Drop-Off
This is where the analysis surfaces something that no amount of dashboard-glancing would reveal. The onboarding sequence has 6 emails, and the engagement curve shows a cliff:
Emails 1 through 3 maintain strong engagement with 35-40% open rates. Email 4 drops to 14% -- a 62% decline in a single step. Emails 5 and 6 never recover, hovering around 12-15%.
Looking at the content: email 4 links to a setup page that requires 8 configuration steps, including API key generation, webhook setup, and a test integration. New users hit a wall. They abandon the setup, stop opening emails, and a significant percentage never activate. The fix is not a better email -- it is a simpler setup page. That insight would never surface from looking at open rates one campaign at a time, because each individual email's rate looks like just one data point. The pattern only emerges when you see the sequence as a whole.
Step 5: Generate the Full Report
The comprehensive report (email-campaign-analysis-h1-2026.md) distills everything into five actionable recommendations:
- Shift send schedule to Tuesday/Thursday at 9 AM -- projected +26% opens based on historical day/time patterns
- Use question-based subject lines more frequently -- they outperform every other pattern by a wide margin, and the team has been defaulting to declarative statements
- Fix onboarding email 4 -- the 62% drop-off between emails 3 and 4 is bleeding new user activation. Simplify the setup page it links to.
- Sunset the monthly newsletter format -- lowest engagement across all campaign types, highest unsubscribe rate, and the most time-consuming to produce
- A/B test personalized vs. non-personalized subject lines systematically -- the data shows personalization works, but a controlled test would quantify the lift precisely
Real-World Example
Marco is the growth marketer at a 20-person project management SaaS. He sends campaigns weekly but only glances at open rates in the email dashboard -- there is never time to dig into what is actually driving performance across months of sends.
He exports 6 months of campaign CSVs and runs the full analysis on a Monday morning. The subject line findings alone change his approach: question-based subject lines outperform everything else by 28%, and he had been defaulting to declarative statements like "New Feature: Timeline View" when "Have you tried the new Timeline View?" would have performed dramatically better. The send time analysis shows Tuesday at 9 AM outperforms his usual Friday afternoon sends by nearly 70% on click rate.
But the biggest discovery is the onboarding sequence. Email 4's 62% drop-off traces back to a confusing setup page, not a bad email. Marco works with engineering to simplify the setup flow from 8 steps to 3, restructures his send schedule to Tuesday/Thursday mornings, and shifts to question-based subjects. Over the next month, open rates climb from 24% to 31%, click-through rates nearly double, and onboarding completion rates jump 40%. The data was always there in those CSV exports -- it just needed someone to look at all of it together.