Running Growth Experiments to Expand Reach for a Prominent Podcast Show

I began by analyzing historical download data, promotional platforms, engagement, and the existing marketing strategy. I found LinkedIn generated the highest engagement, and Elle’s personal LinkedIn account consistently performed best. I also spoke with Elle to clarify the podcast’s goals and asked whether changes to episode content or paid advertising were being considered.

I then researched B2B podcast growth best practices and found that successful podcasts often promote each episode three to four times after launch, using varied content to avoid repetition. Based on these findings, I hypothesized that increasing post-launch promotion could extend an episode’s reach beyond release day and increase downloads.

For each episode, I implemented a schedule consisting of one pre-launch post, one launch-day post, and one to two post-launch posts. I kept the podcast content consistent and focused the experiment on promotional frequency and content variety. To preserve the test's validity, I continued collecting data rather than changing multiple variables too quickly.

I used Transistor as the primary measurement source because it brought together the podcast’s Spotify and Apple Podcasts listening data, while YouTube provided a separate view of performance for long-form episodes and Shorts. I used these channels to compare how different promotional frequencies and content formats affected discovery beyond launch day.

To test content variation without making every asset from scratch, I built an AI workflow using a custom Codex skill. I supplied the workflow with an episode transcript, and it identified the central playbook—usually three to five practical steps—before turning that structure into a repeatable carousel. I also configured it to use Product Marketing Adventures’ existing designs and themes, so the generated content was consistent with the show’s brand rather than looking like generic AI output. The workflow typically saved me around 30–60 minutes per carousel and made it more practical to create multiple post-launch assets for each episode.

The most valuable tools were the analytics sources and the repeatable AI workflow together: the data showed me what was happening across channels, while the workflow reduced the effort required to produce varied follow-up content. I would use the workflow again, but I would still review and edit each carousel for accuracy, clarity, and brand fit before publishing.

The experiment is still running, but the early results support the hypothesis that continued post-launch promotion can drive additional listeners and viewers after release day. Downloads increased by 16% during Days 2–7 after each episode was released, while first-week downloads increased by 4%.

During the experiment period, June 9–July 20, YouTube performance improved across nearly every key metric compared with the previous equal-length period, April 28–June 8. Total channel views increased 29%, from 534 to 690. Long-form views rose 24%, from 235 to 292, while Shorts views increased 33%, from 299 to 397. Long-form discoverability also improved: impressions increased 25%, from 3.6K to 4.5K, and click-through rate rose from 2.6% to 3.1%. Shorts generated one new subscriber, compared with none during the previous period.

These are directional results rather than a final causal measurement, since the experiment is ongoing and other factors may have contributed. One of my biggest learnings was that growth experiments require patience: meaningful results need enough time to emerge. I also learned to combine quantitative data with qualitative research—analytics showed me the patterns, but understanding the business goals and researching industry practices helped me decide what to test. Next time, I would define clearer success metrics and a measurement plan before launch, while continuing to test one major variable at a time.

Check out my site, where I have the experiment, analysis, and supporting visuals in Growth Experiment.