Dot Plots: How to Actually See What Your Users Are Doing

Predictions and Expectations:

• The speaker believes that if founders only look at aggregate metrics like DAUs, they will fail to understand whether users are actually enjoying the product.
• The speaker expects that using dot plots will allow teams to notice patterns they would never figure out "a priori" on their own.
• The speaker predicts that if founders look at dot plots instead of just DAU graphs, they will gain a much richer understanding of user behavior, potentially inferring usage contexts like "office" or "work week."
• The speaker expects that if a company looks at dot plots, they could have identified that a specific contract was "in jeopardy" before the customer churned.
• The speaker believes that modern AI coding tools can build a dot plot visualization tool in "like 10 minutes."

Timelines and Milestones:

• The speaker references hearing the idea of dot plots "10 years ago" from Max Levchin at PayPal.
• The speaker mentions a specific event occurred when a champion at a customer company "left" and a new person came in to "churn," implying an immediate reaction to the departure.

Technology and Product Direction:

• The speaker plans for users to encode states like device type (iOS/Android) or location with "other symbols or shading the cells different colors."
• The speaker suggests that users can sort rows by attributes, such as only looking at iOS users or users whose first day was a specific Monday.
• The speaker recommends that companies sort dot plots based on specific user attributes to "zoom out" and see aggregate patterns.
• The speaker expects that teams can "change the dots to be different symbols" to represent specific features, such as adding an 's' for search or a 'p' for joining a public playlist.

Market and Industry Outlook:

• The speaker expects dot plots to be applicable to B2B products, even though founders might assume they only matter for products that sell seats to businesses.
• The speaker states that the tool scales from a "very small number of users" at the beginning to when a company has "thousands or millions or billions of users."

Company Plans:

• The speaker recalls using dot plots at Google Photos when the company had "well more than a billion users" to sample and analyze different user groups on printed paper.
• The speaker mentions a specific recent YC batch company that signed an "$80,000 a year contract" for 10 seats but only saw 3 seats activate.

Financial Guidance:

• The speaker cites a specific example of a company that signed an "$80,000 a year contract" which later churned due to low activation and sporadic usage.

Risks and Caveats:

• The speaker warns that the number one mistake is charting the "wrong event," such as "open the app" or "signed into the product," which might feel good but doesn't measure real value.
• The speaker cautions that picking a time period that is "too wide," like weeks instead of days, makes it "way harder to figure out what's actually going on."
• The speaker notes that without dot plots, founders "would have no idea" about specific user segments, such as those who only use the product on weekends versus weekdays.
• The speaker implies a risk that aggregate graphs "tend to be going up and to the right, even if users aren't actually enjoying using your product."

Confidence and Disagreement:

• The speaker states with confidence that cohort retention curves and dot plots are "two of the most important tools" for understanding users.
• The speaker suggests that until a company has "hundreds of users," the dot plot "could be your only dashboard."
• The speaker asserts that dot plots provide the "color to go ask the right questions" and build features that you "would never learn by looking at aggregate metrics."