An Open Letter to Pied Piper: Why DAUs are DAUseless
Recently, Rex Sorgatz wrote an article for Backchannel on the real-life models for HBO’s Silicon Valley. And during last night’s episode, my heart sunk as I realized who the joke was on this time: Pretty much everyone in the Valley.
ICYMI, the episode opens with Pied Piper celebrating as its 500,000th user installs the app. But Richard and Jared know the truth: Their Daily Active Users (DAU) stands at a very underwhelming 19,000.
What does that mean? F0r all of the people who’ve downloaded the Pied Piper app, only 3.8% are using it daily.
In tech, that’s a bad sign of health. Or, as Richard puts it to Monica:
“Installs isn’t really the metric that matters […] We need daily active users.”
He’s right that Pied Piper needs daily active users, but in another sense Richard’s also wrong. The scene suggests that DAU is the metric that matters. As Gilfoyle would say: Lie.
And yet a lot of people in the Valley cling to DAU. It’s a popular vanity metric. And so while DAU may be useful to investors, like Laurie Bream, it’s useless to product people, like Richard. Here’s why.
DAUseless
When Pied Piper first discovers that only 3.8% of its adopters are daily active users, the reaction is this:
“Okay, we use that money to do outreach, tutorials, user meetups, uh, create a guerrilla marketing campaign that explains how cool the platform is.”
No.
No, no, no, no, no, no, no.
None of these solutions are attacking the problem, because the problem isn’t actually DAU. DAU’s the smoke, but it sure as hell ain’t the fire.
So, what is the fire?
I’ll answer that question in a minute, but first, what on earth is this?
Is that supposed to be a dashboard? Because I’m not sure what I’m supposed to do with that. It’s essentially just a counter. Watching this, there’s no actionable steps to be taken.
So we ditched that and and implemented Mixpanel into Pied Piper:
There’s our Pied Piper funnel from after they install the app to engagement. Using this dashboard, we can trace users from when they create their accounts to when they share a file. In this walkthrough, we should be able to isolate where the drop-off occurs.
Let’s take it step-by-step.
First off, we know over 500,000 people have installed the Pied Piper app, and looking at our funnel, we see there are 500,008 account creations. Excellent!
And of those 500,008?
83.9% continue to upload a file. That’s a pretty strong conversion, and the market research backs it up. In the focus group, Bernice didn’t have trouble uploading her file.
Next up in the funnel is the “View File” step. “View File” fires when a user is able to review the files they’ve uploaded. Let’s take a look…
And wow. Not so good. That’s almost as ugly as Jared’s jacket.
Okay, the jacket’s still worse. At least with the funnel, there’s possibility for improvement.
Now we know that “View File” is the step in the funnel where Pied Piper is losing the majority of its users. This drop-off also corroborates what the market research group says: Uploading files is easy, but viewing them is difficult and unintuitive.
In a pool of 500,000, 21,023 is pretty close to 19,000. The 21,023 performing the “View File” action are, of course, not directly equivalent to the 19,000 daily active users (not all of them take the same action). But because of their overlap, we can infer that there is a high correlation between the “View File” step and whether or not someone becomes a DAU. If we wanted to track users’ sessions to confirm this hypothesis, we could do that too through People profiles.
So that number (and not DAU) is where Richard’s energy should be focused.
Post-morterm
What can Pied Piper do to save itself?
Instead of explaining that eggs are made out of electrons, the funnel has pinpointed a much more precise step where users needs to be educated: How do I view my files?
Without optimization around that step, users are likely to churn.
Market research is more expensive than analytics, but that’s not even their worst mistake. Their worst mistake is that is all of the episode’s solutions are trying to affect DAU, which is an output, instead of user experience, which is an input. The closest the team comes to addressing their actual problem is (unfortunately) Pipey the Pied Piper Piper.
A lot of startups prioritize big picture metrics without thinking about the meaningful metrics behind those. It’s an okay strategy to woo investors, but a poor way to build products.
Last summer, Periscope wrote up a summary of its user data and took this mentality to task:
Optimizing for DAU/MAU doesn’t properly motivate our team to create a product that people love. Here’s why: if we were motivated to grow DAU, we’d be incentivized to invest in a host of conventional growth hacks, viral mechanics, and marketing to drive up downloads. This direction doesn’t necessarily lead to a better product, or lead to success for Periscopers. We hold ourselves accountable to Time Watched as an organizational measure because it reflects the kernel of our product, and our core values.
It’s not an “if you build it, they will come” thing. Users are already coming to Pied Piper; they’re just leaving, because they’re having a shitty experience.
I don’t know what’s going to happen when the Bangladeshi click farm is exposed, but if Richard wants to see more daily active users, he needs to figure out how to optimize his users’ experience at the relevant part in the flow.
Somehow I can’t imagine this being next week’s finale, but if Mike Judge changes his mind, he knows where to find me.
To read about startups that use data with more success than Pied Piper, head over to The Signal.
