Almost every team I talk to says "Our data is mess!"
@crystalwidjaja (Fmr SVP BI & Growth at Gojek and @reforge Partner) wrote an excellent piece on the real root causes of analytics failures, and a step by step process on how she thinks about it: reforge.com/blog/why-most-…
🧵👇
First, if you don't know @crystalwidjaja and the story of Gojek, it's impressive. Crystal joined at 30 people and helped them scale to complete more daily food orders than Grubhub, Uber Eats, and DoorDash combined, and more trips than Lyft per day 🤯
There are a lot of symptoms of bad analytics/data:
1. Lack of shared language 2. Slow transfer of knowledge 3. Lack of trust 4. Inability to act on data quickly
...
...but these are the symptoms. The real root causes of these symptoms tend to come down to one or more things:
1. Tracking metrics vs analyzing metrics as the goal. Those two things are very different, with the latter being making information actionable.
...
2. Having a developer/data mindset vs business user mindset. A core principle of building any good product is deeply understanding your target user. When building data systems many teams lose sight of who their customer is, or don't have one in mind at all - the business user.
3. Wrong level of abstraction. Bad tracking is when our events are too broad, good tracking is when our events are too specific, great tracking is when we have balanced the two. (example in the full post)
4. Written Only vs Visual communication. No matter how well we name or define events and properties, nothing is more clear than a visual that corresponds with an event.
5. The good old data wheel of death.
So what to do instead? Crystal provided a step by step on how she thinks through the instrumentation process.
A. To get to the right level of abstraction, you need to track journey's of intent, success, and failure NOT metrics.
B. One of the most commonly missed, but important properties on events are contextual properties to help understand what might influence user motivation. For example, how many drivers on the screen when booking in a ride-sharing app?
C. Pressure test your analytics plan with three audiences. Business users close to product development, biz users farther away from product dev (i.e. customer support) and new employees. All have different levels of knowledge/perspective.
D. Keep a "decisions made without data" sheet every quarter to help uncover things you overlooked or didn't anticipate.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
The Entertainment Value Curve - Awesome post by @ravi_mehta (Former CPO Tinder, FB, TripAdvisor) on product strategy in the social space and why TikTok is on 🔥 and Quibi is 📉
Word of Mouth is critical, but notoriously hard to measure and therefore hard to influence. @ybhaijee (Former VP Growth @ Eaze) wrote an excellent post on @reforge about The Word of Mouth Coefficient with some analysis: reforge.com/blog/word-of-m…
Full Thread 👇👇👇
When @ybhaijee and @tomaspueyo worked at Zynga together, they wanted a metric for Word of Mouth that was:
1. Based on Active Users 2. Stable enough to be used in forecasting 3. Could be influenced with product/marketing initiatives
The result was the Word of Mouth Coefficient
WOM Coefficient: Says that for every X active user, you will bet Y new organic users in that time period.
One key 🔑 is that rather than basing the metric on new users (like K-Factor) they based it on active users. Retention is at the foundation of every growth loop...
All product work is not equal. There is a common issue of over-applying one process, measure of success, and strategy to all product problems. @far33d and @onecaseman wrote a monster post talking that is well worth the read -> reforge.com/blog/product-w…
Full thread 👇
"A common conflict I've seen is when product leaders try to apply a single process to all product work...growth and feature work are different and energy is wasted trying to force-fit into the same process, success metric, and approach." - @far33d
There are four types of product work beyond product-market fit:
1. Feature Work 2. Growth Work 3. Scaling Work 4. Product Market Fit Expansion
"What got you here, won't get you there." @onecaseman and @far33d broke down the transition from Product Manager to Product Leader. Excellent insights from @iambangaly and @ravi_mehta as well.
It seems we are shifting into a new phase of this new world. @far33d and I wrote a piece on how to think about your retention strategy should shift with it (w/ contributions from @iambangaly@ElenaVerna@danhockenmaier )👇
There will be different strategies depending on where you sit on the headwind/tailwind spectrum.
Extreme Headwinds = habit has been broken w/ majority of your customers.
Extreme Tailwinds = habit has been accelerated w/majority of target customers and new audiences.
No matter where you are at on the spectrum, don't forget the principles of retention:
1. Retention is about building and deepening habits. 2. Retention is about usage, not revenue. 3. Retention is an output. To move retention, you need to focus on an input.
2/ The MOST important question your team should be able to answer is: "How does your product grow?" This seems like a simple question, but you'll typically find everyone on your team has a different and/or incomplete answer...
3/ This is a really big problem. If everyone has a different/incomplete picture of how the product grows, then you can't have apples to apples discussions about priorities, metrics, goals, or strategy. This leads to people moving in opposite directions.