This is my first crack at making a forecast as a part of what I’m calling the “Cassandra Project.” The gist of the motivation for the project is that product management and entrepreneurship requires smart bets on the future and that you can’t get good at making smart bets on the future without practice. You can read more here.
The forecast I’m making relates to this question:
Will the percent change in iPhone revenue growth from Q1 2018 to Q1 2019 be greater than the percent change in revenue growth from Q4 2017 to Q4 2018?
Base rate
One of the things that supposedly helps with forecasting is finding a base rate, and — in cases of high uncertainty — not straying to far from that base rate. (This is covered in both Superforecasting and in Thinking Fast and Slow) Given Apple’s last quarter and the uncertainty around whether apple is at a sort of inflection point with iPhone growth, this feels like good advice.
Looking at Apple’s 10-Qs from 2018 to 2013, here’s what we see:
| year | Q4 | Q1 | iPhone revenue growth better? |
|---|---|---|---|
| 2017 - 2018 | 14 | 13 | yes |
| 2016 - 2017 | 1 | 5 | no |
| 2015 - 2016 | -18 | 1 | no |
| 2014 - 2015 | 55 | 57 | no |
| 2013 - 2014 | 14 | 6 | yes |
| 2012 - 2013 | 3 | 28 | no |
I stopped at 2013 because prior to that time the iPhone revenue wasn’t called out as as standalone item in the 10-Qs (it was bundled with other services). More often than not, Q1 isn’t better than Q4 in terms of revenue growth. More specifically, only 1/3 times is it the case that Q1 is better than Q4, so that’ll be our base rate.
Fermian decomposition
Another thing that supposedly helps with forecasting is to “fermi-ize a problem,” or as Tetlock put more clearly in the 10 commandments for forecasters to “Break seemingly intractable problems into tractable sub-problems.” Taking that approach, here’s what I’ve come up with:
The # of iPhones sold is driven by:
- population growth (more people means more phones)
- shifting cultural norms (more acceptance of smart phones means a larger percentage of the population will own smartphones)
- alternatives (google pixel is a decent substitute to an iPhone)
- new features that existing users want
There’s not any massive changes with (1), so I’m ignoring it as a potential driver. Same with (2) and (4). My impression of the features on the iphone XS and XR are “meh.” That could be wrong of course, but that’s my impression.
(3) actually matters here. Because of the trade issues /w China etc., Chinese smart phone manufacturers have taken a bite out of Apple’s market share in China. This makes it less likely that their revenue growth will be better than last quarter.
Forecast
So, with all of this, I’d say there’s about a 25% chance that Q1 iphone revenue growth beats Q4 iphone revenue growth. I’m adjusting downward from 33% because of (3) above. We’ll see if I’m right in a week or two.
I’ve also updated the forecast here.