How to quickly and cheaply filter out ideas that won't help you.
“At the end of the day, very few companies fail because their technology doesn’t work. Companies fail for lack of customers.”
Lean Startup “is focused on helping companies be successful by quickly and iteratively determining what customers want, building something that fills only that need, validating the solution, and repeating.”
Next, how do you achieve the high-level goal with an actual product? Are you going to iterate an existing product or build a new product?
Quantitative
“Quantitative sources are important because they allow us to collect data on how people actually use our product, to use that data to find insights, and then to apply those insights to determine what to do next.”
Lyft Example
“For example, perhaps Lyft wants to increase customer retention by getting people to use Lyft instead of Uber. We can turn that into a specific product goal by focusing on how to decrease the number of rides cancelled—fewer cancellations means more passengers served and happier drivers, which means more payment. This will also come by tweaking the current product, rather than building something new. When you cancel a ride in Lyft, you need to specify a reason. A PM at Lyft could look at the responses and come up with ideas about how to mitigate those reasons. Maybe he notices that midday the number of people cancelling their rides due to wait times skyrockets compared to the morning or evening. As a result, perhaps he’d find ways to incentivize drivers to be on the road midday, decreasing wait times, decreasing cancellations, and ultimately keeping customers in a Lyft rather than an Uber.”
Tools to use
After segmenting, if metric is not inline with baseline expectation, flag it to focus on.
Cohort Analysis - similar to segmentation but uses a point in time as a key characteristic. Compare cohorts after time, 2 months, 6 months, etc.
Flag metric when a substantial difference in behavior at a certain point.
Funnel - measure key steps along a user's journey towards some task and group the m together, in journey order.
Many people will complete the first step but not as many complete the last step
Ex. Shopping cart abandonment
Flag metric when/where there is a significant drop off
Hooked Method
Pinterest example of hooked method
Gamification - adding game type features for users - example: Kahn Academy
Excerpts From: Josh Anon. “The Product Book.”
“At the end of the day, very few companies fail because their technology doesn’t work. Companies fail for lack of customers.”
“in a startup no facts exist inside the building; only opinions.” ” - Steve Blank
Lean Startup “is focused on helping companies be successful by quickly and iteratively determining what customers want, building something that fills only that need, validating the solution, and repeating.”
WHAT'S YOUR GOAL AND HOW DO YOU WANT TO ACHIEVE IT?
The very first thing to do is establish your goal for the next product development lifecycle. Examples include, new customers, improve revenue, growth, customer satisfaction?Next, how do you achieve the high-level goal with an actual product? Are you going to iterate an existing product or build a new product?
Iterating an existing Product
- Translate high-level business goals into a product goal
- Focus on what changes can be made to the product to achieve the goal(s)
Example, goal is to improve customer satisfaction. Get users to read more articles because that's a sign they like the content.
Another example, when Facebook changed the Like button to include other reactions,
Note, if you do not have product/market fit, focus on achieving it before revenue or growth goals.
Build a new Product
Building a new product is when you need to make a big transformation to address a need that might not exist yet but you believe will exist in the future.
QUANTITATIVELY FINDING AN OPPORTUNITY HYPOTHESIS
There are two major ways to determine what we want to do to achieve our goal(s).
- Qualitative Reasoning - “involves more abstract concepts like looking at your overall product vision or your intuition based on your knowledge of your customers in determining the next step to take.”
- Quantitative Reasoning - “involves looking at data, interpreting it, and using that analysis to determine what to do next”
Quantitative
“Quantitative sources are important because they allow us to collect data on how people actually use our product, to use that data to find insights, and then to apply those insights to determine what to do next.”
Lyft Example
“For example, perhaps Lyft wants to increase customer retention by getting people to use Lyft instead of Uber. We can turn that into a specific product goal by focusing on how to decrease the number of rides cancelled—fewer cancellations means more passengers served and happier drivers, which means more payment. This will also come by tweaking the current product, rather than building something new. When you cancel a ride in Lyft, you need to specify a reason. A PM at Lyft could look at the responses and come up with ideas about how to mitigate those reasons. Maybe he notices that midday the number of people cancelling their rides due to wait times skyrockets compared to the morning or evening. As a result, perhaps he’d find ways to incentivize drivers to be on the road midday, decreasing wait times, decreasing cancellations, and ultimately keeping customers in a Lyft rather than an Uber.”
Metrics and Analytics
- Collect Metrics
- Determine which one's are success metrics
- Analyze Success Metrics
- Analytics is the process of gathering and analyzing metrics collected.
Tools to use
- Google Analytics
- Mix Panel
Breaking Down Analytics
- Double check Key Success Metrics support the goal
- Assess all data points you are recording, remove irrelevant ones, before making decisions on data.
- Focus on specific success metrics and supporting metrics
- Group metrics together to spot trends and opportunities
- 3 key ways Segmentation, Cohort Analysis, Funnels
After segmenting, if metric is not inline with baseline expectation, flag it to focus on.
Cohort Analysis - similar to segmentation but uses a point in time as a key characteristic. Compare cohorts after time, 2 months, 6 months, etc.
Flag metric when a substantial difference in behavior at a certain point.
Funnel - measure key steps along a user's journey towards some task and group the m together, in journey order.
Many people will complete the first step but not as many complete the last step
Ex. Shopping cart abandonment
Flag metric when/where there is a significant drop off
Turning Metrics into Opportunities by Asking Why
- Once you have metrics, it tells you what is happening but not why.
- When you have a metic that isn't where you want it to be ask why, if the answer is not a hypothesis, keep asking why until you get to a specific hypothesis.
- If a funnel, look up the funnel for "leakage"
- Segment audience to understand how each segment performs or if perform differently
“Here are a few common, interesting opportunities you’ll find for mobile and web apps in metrics:
- A low time on a screen and a high bounce rate—people who leave after viewing this content—on a page that’s supposed to be important likely indicates a mismatch between expectations and reality. The content wasn’t what the customers expected, so they left.
- A long time on a screen and a high bounce rate could be fine if the page is a long article, but if it’s a page with very little content but lots of links, it could indicate the screen is unclear.
- A high number of screen views could indicate that this part of the app is important, therefore you need to optimize it well.
- A low number of views could mean this section is hard to find.”
Intercom's Feature Audit
Hooked Method
Pinterest example of hooked method
Gamification - adding game type features for users - example: Kahn Academy
Surveys
Can use NPS, open ended surveys for feedback. Tools like Intercom.io, delighted to easily integrate.
Customer Interviews
Should not be used for first opportunity hypothesis. Start with an opportunity hypothesis, even if vague, then use customer interviews to validate the hypothesis.Excerpts From: Josh Anon. “The Product Book.”
Comments
Post a Comment