Cross-Platform Rank-and-Join Analysis
Normalize platform rankings, join the results, and find shared winners
- Difficulty
- Moderate
- Time to result
- ~days to results
- Steps
- 7
- Confidence
- 98%
Cross-Platform Rank-and-Join Analysis makes unlike platform metrics comparable without pretending they are identical. Export one file per platform, inspect its columns, and rank the same content independently using each platform's native success measure, such as downloads or views. Join the ranked tables through a shared identifier such as publication date or episode ID. Then calculate rank gaps, isolate records that perform similarly, and visualize the overall correlation. Strong alignment identifies content with broad appeal; large gaps expose platform-specific preferences. The method turns separate analytics exports into a unified view that supports programming, packaging, and distribution decisions while preserving the meaning of each platform's original metric.
Origin
Extracted from Marketing Against The Grain. Kieran Flanagan demonstrated the method by combining 2023 RSS-download and YouTube-view exports for the show, then examining rank alignment and correlation.
Core principles
- 01Compare unlike platforms through normalized ranks
- 02Preserve each platform's native success metric
- 03Join records through a reliable shared key
- 04Investigate both alignment and divergence
- 05Use visualization to reveal relationships
How to run it
- 1
Export platform data
Obtain one raw file for each platform, retaining identifiers, publication dates, titles, and the platform's primary performance metric.
Pro tip Use the same reporting period for every export.
Watch out Mismatched date ranges will distort the comparison.
- 2
Inspect the files
Ask the analytical tool to describe each file's columns, types, missing values, and likely identifiers before manipulating it.
Pro tip Confirm metric definitions as well as column names.
Watch out Do not assume similarly named metrics measure the same behavior.
- 3
Create independent rankings
Sort each platform's records by its native success metric and assign positions from best to worst.
Pro tip Retain both the rank and raw metric for later interpretation.
Watch out Do not combine raw views and downloads directly as though their scales were equivalent.
- 4
Join matching content
Merge the ranked tables using a reliable shared field such as episode ID or publication date.
Pro tip Prefer a unique content identifier when one exists.
Watch out Date-based joins can mispair records if several items were published on the same day.
- 5
Measure rank alignment
Calculate the distance between platform ranks and identify content whose positions fall within a chosen threshold.
Pro tip Review several thresholds rather than treating one cutoff as universal.
Watch out Similar ranks do not establish why the content succeeded.
- 6
Visualize correlation
Create a scatterplot or rank comparison chart and quantify whether performance moves together across platforms.
Pro tip Label major outliers so their topics and packaging can be inspected.
Watch out A weak correlation does not mean either platform's data is wrong.
- 7
Form content hypotheses
Study shared winners, platform-specific winners, and outliers to propose topics or formats for future tests.
Pro tip Treat generated recommendations as starting points for editorial judgment.
Watch out Do not infer a causal formula from a small number of successful episodes.
In the wild
The hosts ranked 2023 episodes by RSS downloads and separately by YouTube views, joined the tables by publication date, and searched for episodes whose positions were within five places of each other. Only three episodes met that condition, while the visualization showed little overall correlation.
→ They identified the need to discover topics in the overlap between their RSS and YouTube audiences.
A B2B publisher ranks articles by newsletter click-through rate and LinkedIn engagement, joins matching articles by content ID, and examines rank gaps. Shared winners become candidates for deeper reports, while LinkedIn-only winners inform social packaging tests.
→ The publisher develops separate platform tactics while identifying themes with cross-channel potential.
Common mistakes
Comparing raw totals directly
Views and downloads can operate on incomparable scales. Independent rankings create a common positional representation without erasing the native metrics.
Joining on an unreliable key
Incorrect matches contaminate every downstream chart and conclusion. Validate the merged records before interpreting them.
Confusing correlation with a content formula
Rank alignment indicates a relationship, not the cause of success. Use it to generate hypotheses for subsequent tests.
Is it for you?
Best for
It is best for creators and marketing teams distributing equivalent content through two or more platforms.
Not ideal for
It is not ideal when records cannot be matched reliably or when the platforms contain fundamentally different content.
From the transcript
“Created a list from one to the bottom, stack ranked on YouTube views, and then start to like blend those two tables together so we…”
“And then ask Code Interpreter based upon the date that was published to join those two things together”
“show me the episodes that performed within five positions of each other.”
From the episode
ChatGPT’s Code Interpreter: Your $20 Personal Data Analyst (#141)