Number Games

Chromecast has been used 400,000,000 times. What does that mean?

[update: I wrote a foll0w-up a year later. Find it here.]

Many years ago, I read A Mathematician Reads the Newspaper. It had a simple, unsurprising premise: numbers can be manipulated to drive people to whatever conclusion you want. It was interesting seeing all the clever ways people use to deputize numbers into service of a greater story.

Amazon is great at this. They skate by with vague statements like “the highest selling product in the electronics category” or “the fastest selling reader in fiscal quarter Q2.” We hear “biggest” or “fastest” and think “sounds like Amazon is successful” before clicking the next article. But Amazon rarely releases actual numbers. And why should they?

Lately we’ve been seeing a lot of “shipped” versus “sold” numbers. Selling a product actually means a successful sell-through. “Shipping” a product lately means you stuffed the channel and product is unsold. But when you report something like “Samsung ships a record x million devices”, the average reader thinks “sounds like Samsung is successful”. Then they click the next article. (When, in fact, Samsung is struggling right now.)

Which brings me to Chromecast. All Google will say is they’ve sold “millions” of the $35, (presumably) break-even device. But recently they announced 400 million “sessions”. Sounds impressive! A recent headline states “Chromecast turns one: why this small streaming stick became such a big deal” and the subheads are “So cheap, and so different”, “400 million cast sessions”, “Competitors are getting the streaming stick fever”, and “Why Chromecast continues to be disruptive”.

So kudos to Google for an enormous number, and for getting great press from it. But, wait. We’re actually going to record “uses” of products now? Well, sure. Because it makes the number look bigger. Why wouldn’t Google report a number that makes you shrug and think “Google seems to be doing well” before reading the next article? Or buying into what’s clearly a popular device backed by what sounds like a stable ecosystem?

I ran some simple math to tease out some numbers a bit more.

If you own an Apple TV, Roku, Chromecast, or other similar device, you’re using it fairly frequently. I’d assume anywhere from 4 to 7 uses per week in a household is probably normal. Let’s go with that.

400,000,000 uses per year, if every user uses it seven days a week on average, means 1,095,890 customers globally. Whereas if each user is using it four times a week, that’s 208 uses per year. That would point to 1,923,077 customers.

(a side note: switching from Hulu to Netflix counts as two sessions. Disconnecting and having to re-connect is two sessions as well. For these numbers, I’m ignoring these situations, but they’re certainly not uncommon, and would inflate the numbers.)

Since Google has said they’ve sold “millions”, 1.9 is a bit low. So let’s adjust usage down a little bit to get past 2 million. That gives us a guess of 2,564,103 Chromecast users using the device 3 times a week.

What about Apple?

In April, Apple said it has sold 20 million Apple TVs total, and the current numbers seem to point to 8-10 million a year. If we assume their engagement is similar to those of Chromecast users (3 times a week), that works out to between 1,248,000,000 and 1,560,000,000 “uses” of the product in the same time period.

Of course, these numbers assume that Chromecast’s piecemeal approach to streaming* drives the same levels of engagement as Apple’s model. There’s evidence that might not be the case. But these are all guesses, of course.

Which is why I try to take these out-of-context, large-sounding numbers with a grain of salt. It’d be helpful if more in the tech press did too.