Image credit: Apple
A Reuters and Ipsos poll of 1,829 Americans showed that 4 percent of American women and 9 percent of American men plan to buy an Apple Watch this year. Using the results and U.S. Census data for demographics, the potential sales number is about 15 million Apple Watches. This is not far from the average of a recent compilation of analyst estimates.
There are various ways to arrive at an estimate of potential Apple Watch sales. One is to develop a methodology. But another is to simply ask potential customers in a random sample from the population as a whole. Of course, polls have to be carefully done.
Ipsos polled 1,829 U.S. adults online between April 8 and 14 about the watch, which opened for preorders last week. The poll had a credibility interval of plus or minus 2.6 percentage points.
If calculated based on 2014 U.S. Census projections, and excluding younger teens, this could mean potential sales of about 15 million watches, if those who said they intended to buy follow through with an actual purchase.
The poll found, not surprisingly, that the highest interest was amongst men aged 18 to 28, an age bracket that had the highest rating of "cool" for the Apple Watch.
Professional pollsters use statistical analysis of a random poll to estimate the probability of error. It can be shown mathematically that a random sample from the entire population can achieve error levels of only a few percent, plus or minus, when the sample size reaches about 1,500 out of several hundred million adults.
A Second Method of Estimation
Another way to approach an estimate is to poll analysts. To varying degrees, financial analysts use estimation tools. For example, one way to start is to look at the population of users who own the required companion iPhone and also have purchase authority as a starting point and go from there. Fortune compiled this data recently and here's what analysts predicted for all Apple Watch sales in 2015.
Image credit: Fortune
A standard technique is to throw out the high and the low numbers, and then take the average. Doing that results in an an average of 20.3 million instead of 22.47 million.
Given that we now have two independent methods of estimation that are not far apart, it's probably a good bet that Apple Watch sales in 2015 will be between 15 million (U.S. polling) and 20 million (analyst average, global).
A third method, Apple's own data on early sales, if published, would be hard to use. That's because the rate of sales due to early adopters likely cannot be sustained over a long period. For example, 975K/day (reported on first day) times 265 sales days in 2015 = 258 million. Not likely. When the early adopters are done, sales could drop off.
Then, perhaps, as the influence of the device percolates in the general population and manufacturing ramps up, sales will pick up again. Only Apple will see this "U" shaped curve in sales reports. Also, recall that Tim Cook has previously announced that Apple isn't planning to report on Apple Watch sales in detail. So we'll have to wait until Apple has enough confidence to post the actual numbers to see how good the estimates have been.