Everybody attention – I’m going to do a forecast at the end of post. It’s a good one. And this post is intentionally picky and personal. Reason is that I see blogging as a form of journalism. And when I see bad journalism, I get picky and personal. But I try to keep it down and start again with a thought game. (Yes, I like them.)
Let’s assume I try to forecast the temperature for next month. My forecast is:
“28 degrees on the first week, 21 on the second. Third week will be 16 degrees and the month ends to quite cold 12 degrees of temperature.“
Now one week passes, 24 degrees. I’ll think it’s close enough. But second week takes me by surprise: 16 degrees! The winter comes faster than I thought! I’ll immediately readjust my forecast to be:
“11 degrees for third week and only 7 degrees for the last week. Then two more weeks of 7 degrees after that.“
Now we get third week in 14 degrees and the last week is 13 degrees – almost double what I expected. The fifth week happens to drop to 8 degrees. What should I say? Well, of course I’ll say:
“I was the most accurate forecaster for how badly temperature would fall. I predicted that temperature would crash to 12 degrees for the last week of this month. Nobody else dared suggest that weather would get so cold, some very cheerfully hoped that weather would still warm up to as high as 28 degrees at the end of month. The reality was.. 12.6 degrees.“
Welcome to my today’s post if you are (like me) shouting: “WHAT ARE YOU TALKING ABOUT? YOUR FORECAST WAS FOR FOUR WEEKS AND SINCE YOUR SECOND WEEK FORECAST MISSED REALITY BY 5 DEGREES YOU GAVE US A NEW ONE THAT MISSED LAST WEEK BY 6 DEGREES! YOU CAN’T GO BACK TO YOUR ORIGINAL FORECAST IF YOU HAVE ALREADY SAID IT’S NO LONGER VALID!”
You have just witnessed a way to be “most accurate forecaster” – invented by Tomi Ahonen.
Yes, Tomi Ahonen is indeed doing some predictions in his (now already second latest) blog. But let’s see how he starts it:
“For new readers – I was the most accurate forecaster for how badly Nokia’s market share would fall after the Elop Effect last year February (the Burning Platforms memo which caused a Ratner Effect on Nokia sales, and the related Microsoft announcement also in February when Nokia had no phones to sell on Windows Phone, which caused an Osborne Effect. I call these two the Elop Effect, the most damaging management communication of all time). I predicted in February 2011, that as Nokia’s new CEO Stephen Elop suddenly and very radically altered Nokia’s long-standing strategy in smartphones, the smartphone sales would stall and Nokia’s market share would crash to 12% by Q4 of 2011 (down from 29% at the time when the forecast was made – this is totally unprecedented in any industry, so there is no comparison to use, no analogy that could help with the forecast). Nobody else dared suggest that Nokia would be so badly damaged, some very cheerfully hoped that Nokia’s market share would be more than twice that, some projected it as high as 28% for the end of the year. The reality was.. 12.6%.” 
You may notice the pun included in my story about temperature. But I want to quote Tomi Ahonen a bit: “I mostly do not comment on forecasts, as forecasts are guesses and your guess is often as good as mine.”  And I want to say this before I start to do some in-depth analysis on his forecasts from the past.
So we know Tomi Ahonen has done three forecasts before the one at hand:
In February 2011 
In July 2011 
And in May 2012 
He had good start, since he was pretty much only one alerting about the Nokia’s record-breaking market loss in 2010 (Yes, you read it right – he was reporting about market share loss in 2010, even though he never mentions it nowadays). But let’s now see how accurate our man has been.
TOMI DOES FORECASTS
Here are all his predictions – timewise – and matching results from Nokia:
Now you can see the reported readings at first column. These are the locked-in facts. I have highlighted with red the predictions of Tomi Ahonen that matches them best. The “reference curve” is list of market share, based on assumption that the “pre-Windows Phone strategy” trend from 2010 (reported by Tomi Ahonen himself) would have continued. Nokia had lost 15% of its market share during Q3 2010. Then 15% of remaining market share in Q4 2010. It’s continuing same math.
Now next column is of course February 2011 forecast of Tomi Ahonen. See how well it fits to the reference curve? Well, we have to assume he used something as a base, no matter how much this is “totally unprecedented in any industry, so there is no comparison to use, no analogy that could help with the forecast“. I admit that his very first forecast is heck of a good forecast – very accurate in long run – done by just extending the current trend. Unfortunately, extending a trend that started before “Elop Effect”.
Now we know that 26th of July 2011 Tomi Ahonen made a new forecast. It’s the next column. Q2 results were bad and he said himself that it was time to change the formula of the model. That prediction went far off – his original forecast was more accurate what comes to BOTH remaining quarters. (More accurate than reference curve, if I may add.)
…So now he refers to the original forecast and says he was “the most accurate forecaster“. As if he never fixed the original one.This is the bad journalism I told about.
There was no new forecast before Q1 2012. But luckily Q1 went really bad, throwing the result to one percent off from Tomi Ahonen’s second forecast. Once again he can say he has been “the most accurate forecaster” – if we pick correct forecast, that is.
At this point of time (1st of May 2012) Tomi hands us his next forecast. I added that to the next column.
And very recently – 27th of June 2012 – we get a third forecast, still covering the same three-quarter period. He now has three forecasts overlapping the same three quarter period.
UPDATE AUGUST 21st 2012:
Tomi released the next market share reading and so far it seems his both adjustments are going wrong. Really interested in seeing the Q3 and Q4 results.
EXNOKIAN DOES FORECAST TOO!
Now it is time to do my forecast, which I assume will be at least as accurate as the one Mr. Ahonen did in February 2011:
- If Nokia’s market share for Q2 2012 is 5 percent or below, Tomi Ahonen will refer to his most recent forecast  and say he was the most accurate forecaster.
- If Nokia’s market share for Q2 2012 is 7 percent or more, Tomi Ahonen will refer to his original forecast from July 2011 .
Furthermore, if Nokia’s market share is 7 or 8 percent, he will in addition say he was the most accurate forecaster.
- If Nokia’s market share for Q2 2012 rounds to 6 percent, Tomi Ahonen will refer to his forecast from May 2012 , say that it is the average of his all three forecasts and say he was the most accurate forecaster.
I think we can wait. Results for Q2 are right behind the corner.
(Okay, I can’t wait – the excitement is killing me!)
UPDATE AUGUST 21st 2012:
Tomi released the next market share readings. He has said in twitter that those are from most accurate forecaster, but unfortunately has not told which one is the most accurate one.
Guideline for commenting: I hate the way Tomi Ahonen deletes criticizing comments from his blog. However, I plan to follow three of his principles: I’ll delete comments that are
- Personal insults to someone
In addition, if you wish to challenge my previous posts, please comment to those.