A DEEP DATA DIVE ON PRODUCT MANAGEMENT ON MEDIUM.COM

by NomNom and DataStories

Imagine you could skim over 4,000 product management articles in less than 10 minutes. Imagine you could easily discover great authors to follow and the content that makes them great without effort.

We've got it for you here.

ABOUT THIS RESEARCH

In the last 2 years we have seen an explosion of content on product management, from blog posts and newsletters to podcasts, and it is easy to feel overwhelmed by the abundance of information available on a large number of publications. We especially noticed a staggering number of posts being published on Medium.com so we decided to take a step back and look into the publishing patterns in the product management content landscape on Medium and see if we could find any metrics that could help us navigate this avalanche of information.

NomNom and DataStories joined forces to analyse over 4,000 posts in the search for trends, hot topics, and influencers.

The research you are about to read is purely based on what the data showed us.

We analysed multiple metrics such as content shares, likes, publishing frequency, as well as lists of all the influencers we could find on the web [Footnote 1]. We hope this research helps you navigate the currents of information available on product management and steers you to new waves worth riding or reefs worth exploring.

To get this report in pdf follow the link and enter $0 as the cost to download:

IN THIS REPORT:

METHODOLOGY

FINDINGS

  1. WHY MEDIUM.COM
  2. CATCH THE WAVE
  3. USER EXPERIENCE IS LEADING THE CHARTS
  4. TOP 17 INFLUENCERS ON MEDIUM
  5. WHAT MEDIUM READERS LIKE
  6. A GROWING TREND
  7. LIKEABILITY ISN'T EVERYTHING
  8. PUBLISHING PATTERNS
  9. DISCOVERING CONTENT ON MEDIUM

FINAL THOUGHTS

METHODOLOGY

In preparation for this research, we looked at ways to retrieve information about product management on the web. We used ahrefs and buzzsumo to identify content with a given keyword (we got 44K posts from ahrefs and 6.7K posts from BuzzSumo Pro). Upon close inspection we discovered that the data was incomplete. It did not capture important blogs and publishers like uxmag.com, Mindtheproduct.com, and svpg.com.

We decided to focus on Medium.com, scrape the data ourselves, and not use any third party tools. We searched and pulled out all available posts with the tag “Product Management” (from here).

Yup, we scraped.

The scraping process consisted of five careful steps:

  1. We retrieved JSON files containing descriptions of all posts tagged as PM.
  2. We pulled out a total of 4759 individual url links from these JSON files.
  3. We cleaned up the links (removed 41 links referring to authors instead of posts, and 2 links to external resources).
  4. We pulled out all contents of each link and created 26 metrics per post (text, title, author, url, number of words, images, videos, sentiments metrics, post date, etc).
  5. We removed all posts with fewer than 100 words in English characters (963 posts were removed).

This got us 3,582 posts with a guaranteed tag of “Product Management”.

Then we dove deeper into the data.

We thoroughly analysed all posts tagged “PM” on Medium. To create this analysis we used DataStories tools and Python.

We assessed the sentiment

We looked at post titles and narratives using the Natural Language Toolkit (NLTK) API from Mashape.

We checked social media response.

After using LinkedIn and Facebook APIs to pull the social shares of the Medium.com posts, we discovered that many posts with a high number of Medium likes had 0 LinkedIn and 0 Facebook shares. Playing it safe, we did not base any conclusions on the social shares stats. We focused on Medium likes.

.

The data we used contained:

  • 3,582 posts on Medium.com
  • published between November 19, 2015, and May 17, 2016
  • in the English language
  • tagged as "Product Management".
  • These posts came from 2,071 unique authors.

We used multiple metrics.

We liked the likes on Medium. But identifying quality content is hard. A single metric approach to filtering content does not always work.

Because Medium.com offers curated content, sorting the posts by the number of Medium likes was reasonable. This approach would not work for data aggregator tools. If you try to identify the top posts, authors, or media by the total number of shares, or the total number of referring domains, or the total number of followers of the author, you will arrive at different and often irrelevant results. Several metrics need to be looked at together to identify genuine producers of consistent and highly-liked content.

We watched the traffic.

If we had the data on how the Medium audience has been evolving over the last 1.5 years, we could establish with certainty whether or not the growing number of posts on PM is related to either a growing interest in Product Management, or is merely a reflection of the growing number of Medium readers and authors. In the absence of this data, we can only check some implicit metrics about the growth of Medium - the traffic.

The ahrefs tools provided the estimation of the traffic to Medium.com since January 2015. See the graph below:

.

The estimated traffic does not display any growth in the first months of 2015, compared with January 2016. This gives us a reason to hypothesize that the increase in the number of posts on PM is an indication that it's the interest for PM that is growing, and not just the number of writers.

We considered post length.

We wanted to see whether post length had been growing over time, as these days super-long, 4,000+ word educational posts are popular marketing tools (e.g. datastories.com/gallery/case-blogs). But on Medium there is no need for tall tales. Short stories get a good buzz.

Only 15 out of 3,500 posts got more than 1,000 likes, and all these posts contain a relatively small number of words (the average number of words in these posts is 1,490 words, while the median is 989 words). For example, the post "Let's talk about Product Management" by Josh Elman has 3,300 likes on Medium (this is the highest number across all PM posts), and only contains 240 words. However it also presents 35 images and one video (see the post here).

Bottom line - in 2016 some posts were super-long. But exceptionally well-liked posts were not that long.

In the graph below we plot all posts by the day they were published. The size of the circle depicts the number of words in the posts (the smallest is 100 words, the biggest has 10,108 words). The y axis also shows the number of likes per post (as of May 17, 2016).

OUR FINDINGS

1 - Why Medium?

Medium = Curation.

Medium's Editors Picks are followed by 625,000 people.

Its democratic ethos and its focus on quality content offer a rich playground for analysis. We're following the traffic, and as Medium's popularity has increased, companies are migrating their blog posts to this organized content publisher.

Our research using different aggregators indicated that Medium offers the most representative sampling of the internet's credible product management content.

Because Medium is curated by both an editorial team and the community, the chances for completely irrelevant posts to be tagged as “Product Management” are small. Additionally, in both our datasets from content data aggregators (ahrefs and BuzzSumo Pro data), Medium.com appeared as the top domain with respect to the number of posts.

Note: As The Marketer's Guide to Medium explains: "The ethos of Medium is inherently democratic; it seeks to give a voice to people who have something interesting to say, even if they don't have thousands of Twitter followers, an active blog or friends in the right places. Medium is built to reward content for its quality, not for the pedigree or popularity of the author....[and] On Medium, the content that is made most visible is not necessarily the most recent, but it is almost certainly the best."

2 - Catch the wave. Product Management content is on the rise.

Look at the number of posts with the tag “Product Management” on Medium.com since the beginning of time up to May 17, 2016:

.

Only 6 months into 2016, product management content has already matched the total number of posts published on the subject in 2015.

3 - User experience is leading the charts.

We analyzed 38 keywords of interest to see the presence they had in all posts, in the top 436 posts, and in the posts of the identified influencers.

To our pleasant surprise, we discovered that User Experience (UX) leads the charts among all posts considered. See the fractions of posts using a particular keyword below:

.

38 Keywords we analyzed:

Out of curiosity we also looked into Google trends and compared customer experience vs user experience and found an upward trend for both terms.

.
Source: Google Trends

This is just another indication of the increasing interest in user experience and customer experience as disciplines. We may be finally entering the customer-obsessed era. After all, according to Walker's research, by 2020 customer experience will overtake price and product as the key brand differentiator.

4 - Top 17 influencers on Medium.

Out of 2,071 unique authors writing on the subject of PM on Medium .com, only 17 authors emerge as influencers according to our definition - people who consistently write unique, highly liked content. This is only 0.8% of authors. Eight of the 17 authors happen to have more than 4,500 followers on Medium.

Most authors either have a lot of posts or comments with few to no likes, or only one or two posts with lots of likes [Footnote 3].

THE TOP MEDIUM.COM INFLUENCERS ON PRODUCT MANAGEMENT ARE:

HOVER over the dates below to see the posts.

Influencer is the one who consistently writes well-liked posts.

Below we show authors as bubbles, with the horizontal axes defining the number of their posts tagged as PM, and the vertical axis depicting the median number of likes. Note, that the average number of likes depends heavily on the outlier posts.

It is hard to consistently write highly liked content.

The top posts with 23 or more likes are so rare that they are statistically considered as anomalies. There are 436 posts, and they constitute 12% of all data. As we examined these posts, we noted that people share their “hearts” sparingly on Medium.

Top Medium Influencers for Product Management and their stats

Author

Median # Likes

# Followers

Average Score

# Posts on Medium

Julie Zhuo 2500.0 73181 8.2 100
Brandon Chu 71.5 1845 7.6 17
Steven Sinofsky 273.0 17751 7.4 33
Jason Fried 156.0 89582 7.3 126
Chandra Kalle 35.0 958 7.2 29
Giff Constable 26.0 2018 7.2 26
Intercom 114.5 16072 6.9 40
Ken Norton 81.5 8204 6.9 20
Matt LeMay 57.0 2054 6.8 12
Josh Elman 346.0 27686 6.8 64
Kimber Lockhart 1359.0 1543 6.7 4
David Cancel 210.0 9692 6.3 45
Paul Adams 232.0 11210 5.9 16
Simon Cross 221.0 2164 5.9 14
Noah Weiss 510.5 4343 5.9 9
Nathan Creswell 422.0 338 5.8 4
Rian Van Der Merwe 255.5 3293 5.6 62

Top words used by Influencers

We also analyzed all posts by the influencers for content and the popular words used (based on the "text frequency - inverse document frequency" , or tf-idf method) and depict them below:

Author

Median # Likes

Top Words

Julie Zhuo 2500.0 [design, designer, pm, id, try, problem, like, good, work, think]
Kimber Lockhart 1359.0 [sense, team, decision, fast, course, make, time, action, day, individual]
Noah Weiss 510.5 [google, search, time, idea, pm, week, ceo, structure, core, engineer]
Nathan Creswell 422.0 [pm, engineer, youre, product, see, something, sound, get, go, thats]
Josh Elman 346.0 [product, twitter, user, facebook, manager, would, turn, talk, hear, company]
Steven Sinofsky 273.0 [product, work, new, change, time, company, best, pm, year, many]
Rian Van Der Merwe 255.5 [research, end, live, product, experience, revenue, need, design, people, didnt]
Paul Adams 232.0 [meet, month, someone, people, hire, pm, need, owner, facebook, build]
Simon Cross 221.0 [pm, book, facebook, trust, read, ive, list, people, post, youre]
David Cancel 210.0 [product, matter, customer, company, team, pm, wrong, get, ceo, co]
Jason Fried 156.0 [work, weve, project, thing, make, time, version, http, client, get]
Intercom 114.5 [customer, product, feedback, first, user, cost, market, business, youre, make]
Ken Norton 81.5 [google, meet, product, pm, hire, engineer, interview, manager, one, ceo]
Brandon Chu 71.5 [product, pm, team, company, launch, user, work, make, get, youre]
Matt LeMay 57.0 [pm, product, skill, work, team, ive, manager, process, customer, often]
Chandra Kalle 35.0 [product, theyre, user, http, push, com, data, mobile, dont, engineer]
Giff Constable 26.0 [team, pm, com, people, talk, usually, thing, part, new, work]

5 - What Medium readers like.

We discovered that having a high number of followers does not correlate to the number of likes authors get for their posts .

Look at the graph below. We would expect the top right corner to be filled with authors (more followers - more likes), but this is not the case at all, except for a single person. Julie Zhuo, Facebook's vice president for product design, has 64K followers, and a median of 2,200 likes for the three posts she wrote for Medium.com!

Having many followers does not guarantee many likes

6 - A growing trend. Expect 4,125 Product Management Posts in 2016.

We observe a consistent growing trend when we display the posts on PM. In the graph below the blue bars represent the actual number of posts in our data, and the green bars are predicted numbers based on the simplest prediction model for linear growth (which has 97% correlation accuracy compared to actual monthly volumes):

There is strong growing trend of the content tagged as product management on medium.com.

So, based on the linear growth and wild predictions over a seven- month period from May to December 2016, we can hypothesize that this year a whopping 4,125 posts will be published in the year of 2016.

7 - Likeability isn't everything. Thought leadership on Medium emerged with few likes.

The median number of likes for all PM posts is only 2. The largest fraction of posts has none to few likes. In fact, 88% of PM posts on Medium get 22 likes or fewer. If a post got more likes than that, it falls among the top 357 outlier posts.

We also looked at the distribution of keywords in the highly liked posts and found it is very similar to the entire universe of PM posts. However, it appears that readers liked the posts with more emphasis on the top keywords. Note the values are higher than in the earlier keywords graphic:

User experience is the main keyword of product management posts on medium.

8 - Publishing patterns: Some Product Management authors find that one post may be enough. Others plunge into posting.

Another interesting fact about top posts is that they have 319 unique authors (only 15% of 2,071 - all authors we considered). The most surprising for us was to see that the majority of these authors (259 people) have only one single post on Medium - this is 81% of authors of "top posts".

At last we looked at yet a smaller, more highly curated sample of Product Management posts only written by people whom we identified as “influencers” based on the number of posts and the median number of likes. In this sample only five keywords were present, with customer experience being the top keyword. (19.5% of posts by "influencers" in our definition use the keyword "customer experience".)

User experience is the main keyword of product management posts on medium.

9 - Discovering content on Medium.

Out of the 883 different tags present along with the tag "Product Management", the top are 'Startup', 'UX', 'Design', 'Tech', 'Agile', and 'Scrum'. Bonus: Try the tool below to quickly retrieve posts that interest you. We built this quick tool to enhance searching by keywords, since we found combining tags the traditional way does not always produce sufficient results.

Type any word here to see top Medium posts with it:

Author

Title

Likes

Date

FINAL THOUGHTS.

We hope this research helped you discover new content and interesting people to follow, and gave you a panoramic view of what's going on at Medium.com when it comes to product management content.

ABOUT NomNom

NomNom is all of your customer feedback and user research in one place. NomNom helps product, UX, and CX teams aggregate, organize, and make sense of customer feedback easily. Try NomNom for free today.

ABOUT DataStories

DataStories is all about letting the data speak for itself. DataStories is a done-for-you analytics platform that helps data owners understand what matters in their data and what to do next through interactive narratives and instant what-if scenarios. Try DataStories with a personal walk-through at platform.datastories.com or call for a custom research report like this one.

If you want to find out more about the companies behind this research, you can find us here:

NomNom Insights. All your customer feedback in one place.

NomNom Insights. All your customer feedback in one place.

DataStories

DataStories. Automated Analytics for busy People.

TAKE THIS RESEARCH WITH YOU

Click the button to download this report in a PDF and enter $0 as the cost to download:

Leave a Comment:

To comment as a guest, please, share your comment, start typing your name and you'll see a check box "I'd rather post as a guest" appear.