Internal-External Campaign Analysis Loop
Turn campaign data and market trends into an evidence-based action calendar.
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 96%
The method gives a reasoning-capable AI model structured information about the product, target audience, campaign objective, and historical results. The model first evaluates each campaign, then searches across the data for recurring relationships between formats, channels, audience groups, timing, conversion, and revenue. It adds an external layer by researching seasonal cycles, industry events, economic conditions, and other factors that may have influenced performance. Those combined internal and external findings become actionable recommendations covering content, channel allocation, budget, timing, and risk. A follow-up prompt then converts the recommendations into a practical calendar with campaign titles, audiences, value propositions, descriptions, and suggested dates. The output remains a decision aid and should be checked against source data and business constraints.
Origin
Extracted from Marketing Against The Grain, where DeepSeek R1 was demonstrated on a fictional SaaS campaign dataset targeting doctors.
Core principles
- 01Combine internal performance data with external context.
- 02Analyze patterns before recommending tactics.
- 03Match campaign formats to audience behavior.
- 04Convert findings into a concrete execution calendar.
- 05Use follow-up prompts to deepen the initial analysis.
How to run it
- 1
Frame the decision
State the product, primary audience, business objective, and the decision the analysis must support. Give the model enough commercial context to interpret the metrics rather than merely restating them.
Pro tip Specify audience subgroups when different buyers may respond to different formats.
Watch out A vague objective will produce broad recommendations that are difficult to act on.
- 2
Supply structured campaign evidence
Provide consistent fields such as dates, budget, audience, channel, conversion rate, revenue, and ROI. Attach the data as an image, file, or spreadsheet when appropriate.
Pro tip Use the same field definitions across every campaign so comparisons remain meaningful.
Watch out Verify that the model has read every value correctly before relying on its analysis.
- 3
Analyze performance and patterns
Ask for campaign-by-campaign evaluation followed by patterns across formats, channels, audiences, and outcomes. Require the model to connect its conclusions to the supplied product and audience details.
Pro tip Ask it to distinguish observations supported by data from hypotheses requiring validation.
Watch out Do not treat correlation between timing or format and performance as proven causation.
- 4
Add external context
Ask the model to research events, seasonal cycles, industry conferences, budget periods, regulatory developments, and economic pressures that could have affected results. Use citations to inspect the external evidence.
Pro tip Name the audience's profession and geography to improve contextual relevance.
Watch out External explanations can become plausible stories unless sources and dates are checked.
- 5
Generate actionable recommendations
Request specific changes to campaign formats, content, channels, budget allocation, timing, and risk mitigation. Tie each recommendation to an internal pattern, an external trend, or both.
Pro tip Require a rationale and expected metric impact for every recommendation.
Watch out Reject generic tactics that are not connected to the analysis.
- 6
Build the execution calendar
Use a follow-up prompt to translate the recommendations into scheduled campaigns with titles, target audiences, value propositions, descriptions, and timing. Review the calendar against capacity and commercial priorities.
Pro tip Ask for dependencies, owners, and success metrics if the plan will move directly into execution.
Watch out A polished calendar is not proof that its assumptions or dates are correct.
In the wild
A SaaS marketer supplies campaign dates, budgets, channels, conversion rates, revenue, product features, and doctor audience segments. The model finds that group practices respond to webinars, hospital administrators favor in-person trust building, and solo practitioners engage poorly with long-form blogs. It combines those findings with annual clinic planning cycles, summer vacations, and medical conference season, then proposes a year-long webinar schedule.
→ Historical performance and external timing signals become a concrete campaign calendar with audience-specific topics and dates.
Common mistakes
Uploading metrics without context
Numbers alone do not explain the product, buyer, objective, or constraints. Include those inputs so the model can interpret performance commercially.
Accepting external claims without checking
Seasonality and event explanations may sound convincing even when weakly sourced. Verify dates, citations, and relevance before changing campaign timing.
Stopping at analysis
Insights create little value if they never become decisions. Use a follow-up prompt to convert validated findings into a scheduled and measurable plan.
Is it for you?
Best for
It is best for B2B marketers with historical campaign data who need to decide what to run, for whom, and when.
Not ideal for
It is not ideal when campaign data is unreliable, objectives are undefined, or recommendations cannot be reviewed by a knowledgeable marketer.
From the transcript
“analyze the performance of each campaign, considering the product details and target audience identify patterns or trends, provide actionable recommendations to optimize future B2B campaigns.”
“also look for external trends I should be mindful of around targeting doctors.”
“I analyzed my campaigns. I got an incredible webinar calendar for the coming year that's mapped to the internal learnings and external trends in two…”
From the episode
DeepSeek R1: The FREE AI That's Beating OpenAI’s o1 Model