The Guide to Product Analytics is here: Behind the scenes
We just launched the first edition of our Guide to Product Analytics, a book of questions and answers from 25+ product leaders at some of the world’s most successful brands. Click here to get full access to the guide, analytics worksheets, and video tutorials that will help your whole team master the foundations of product analytics.
At Mixpanel, we’re building software — and creating content — for product managers, so we approached the writing of this book with a product mindset.
1. First, we did some user interviews.
We spoke with a dozen product managers and data scientists to figure out what kind of content they needed to become more successful with product analytics. Here are a few of the “user stories” we collected:
- “I’d like to know more about best practices in product analytics.”
- “I’d like to see how other companies approach measuring XYZ.”
- “I have a dedicated data team on my project but I’d like to know how I can answer some data questions myself.”
- “There’s not much content these days that’s actionable. I want practical advice.”
We collected all this feedback through calls, emails, and in-person lunch meetings (pre-COVID!) and organized it in a doc.
2. Then we researched the market.
We looked for available resources that solved our user’s problems. Turns out, there weren’t many. Academic books and papers covered the psychology of user behavior and the foundations of quantitative research. Industry blogs had some answers to the questions our users were asking, but didn’t fill in the full picture. Other vendors had written practical how-tos for product analytics, but the advice didn’t come from real PMs solving real problems for different types of products in today’s market.
3. We wrote a spec.
We listed our objectives, goals, features, design notes, risks, etc. For example: We wanted this book to be comprehensive, but not long (50 pages, max). We wanted to add Mixpanel tutorials, but layer them on top of broad industry advice, so that it would be a book about product analytics and not just Mixpanel. We wanted to break it down into “bite-size” content so it would be easy to consume. Most importantly, we wanted to source the best ideas from the best PMs in the industry and share real metrics and stories from the field.
4. We prototyped.
The first prototype was a Google Doc with a long list of questions. We initially called them “recipes” for product analytics; the guide was a “recipe book.” We tested the prototype on a small group of PMs and, after confirming their interest in getting the answers to these questions, we landed on the “book of questions and answers” concept.
5. We sourced content.
We reached out to the brightest PMs and growth experts in the industry, including Josh Elman (VC & Advisor; Former Product Lead at Robinhood), Brian Balfour (CEO, Reforge), Andrew Chen (General Partner at a16z and former Growth at Uber), Lenny Rachitsky (Former Lead PM at Airbnb), Akio Bandle (Sr. PM at ZipRecruiter and a long-time Mixpanel customer), and many others. We also spoke with Mixpanel’s own PMs; for example, our senior PM Moinak Bandyopadhyay contributed to the chapter on engagement analysis.
6. . . .and planned our go-to-market.
Marketing and pricing are important parts of launching a product. (We touch upon monetization lightly in the guide.) As for many content teams, Mixpanel’s price tag for content has historically been a form-fill — we have asked people to give us their email address in exchange for a downloadable PDF. But for this launch, we wanted to open the gate and let users access and browse content freely. It was a bit scary: without a number of “downloads” to show, how will we know we’re successful? After agreeing on a set of metrics to measure engagement, we were ready to get started.
7. We built our MVP.
While research naturally came first, writing and design work were generally happening in parallel. Because we were “ungating” content, we needed to design for the web, not for a PDF. We wanted the content to be searchable, and we wanted to have on-page analytics to measure engagement.
Turns out, when it comes to long-form reading, it’s harder to design for the web. The first version looked nice — but with over 50 pages of content, it felt like an endless click-and-scroll. So we went back to the drawing board and re-structured the book into smaller chunks, each about a page or two long, so readers can click to a specific piece of content and quickly get value from the book without having to read through the whole thing. This meant we needed to make some tradeoffs in style and flow to make usability improvements.
After a couple of iterations, we had the final MVP in Figma, complete and ready for beta testing.
8. We launched it for a small group of users.
Our beta version of the guide was a combination of two Google Docs (the guide and worksheets) and a Figma file, which we sent out to a small group of experts to “test” and review. After a few rounds of feedback, we felt pretty good about the content and design.
9. We QA’d and smashed bugs.
QA is never fun, but it’s critical for any product ready to transition from an MVP to launch-ready.
And so here it is: the final product. It’s the first edition of the Guide to Product Analytics, or v.1 of our product, and I look forward to hearing what you think. Email me directly at firstname.lastname@example.org with any feedback — or if you’d like to be part of v.2!