Thursday, March 18, 2010
History vs. current events
at 2:40 AM
This one shows the whole history of web order data I have, at the finest granularity available (between a week or two since 2005 and perhaps months before 2005). Data from 2001 to 2004 compiled by Steve Evans (aka Caligula). The value charted is the 30 day moving average of the daily rate, on a forward looking basis (i.e. the average for say April 1 to April 30 would be plotted on April 1 along the horizontal axis) in thousands of orders per day.
Here we see the blue line representing the same web order data, but aggregated on a quarterly basis, and also the data for quarterly revenue as reported by Apple (red line) and the ratio of Revenue/Orders (green line). Smoothing was applied because it looks really cool.
As for current events, it seems I've created some sort of controversy by reporting data that showed a steep fallout in sales velocity after the first day of iPad pre-orders, which was of course expected. Maybe some people are better off not knowing precisely at what level and how fast things reach a stable point and would prefer being told a big round number in a month or two, with no interest in the internal dynamics of an event driven launch. A suggestion, don't read these updates if you don't like such fine-grained data analysis, ok? Just wait until my pre-earnings report in mid-April.
It seems this image was a favorite among fellow bloggers. Here it's updated up to the latest order we have, for
yesterday around noon Friday morning.
Now the thing with this image is that it shows the total number of web orders, not just those that include iPads, and there's no way of discriminating these through the factual data. But it's a nice image, and I'm not surprised that people are latching on to it. Now, as shown above I have a rich history of web order data going back years, and have several "sample datasets" for the month of March and for various types of product intros. What I've been doing in this case is assuming the week-long measurement of the rate just before the iPad was available for pre-order has been relatively constant except for a slight increase due to the higher traffic and interest, and subtracted this value from the observed rate.
Doing this definitely introduces some uncertainty, but there's no reason why the non-iPad orders should be getting more or less volume now, except a sensible proportion due to the increased traffic (someone came to the store interested in iPad, saw all the other goodies, and decided NOT to pre-order the iPad but to order something else). Notice that if someone decides to ADD other non-iPad stuff to their iPad pre-order it still counts as an iPad pre-order.
Well, there you have it. That's my update for the week: 180 thousand units pre-ordered online as of noon yesterday. [Update, 190k as of Friday morning, exactly a week after becoming available for pre-orders.] It obviously doesn't include store pickup reservations, nor will it include store sales after April 3 (if I'm still getting order submissions and bothering to track these). That's it. Based on this, and using my intuition and experience, and voodoo or however you want to call it, I'm sticking to my prediction of 1 million sold by mid-April.
As you can see, that conclusion amounts to not much more than my opinion. But the data in the first three graphs is factual, not opinion or estimates. Thus, my recommendation is that you come to your own conclusion as to how this might measure against expectations, both short and long term (although I would advise against putting too much weight into short-term market predictions) and decide for yourself, regardless of whether I'm right or wrong in my opinion-based prediction.
Thanks for all contributions.