New data show Model S sales down 33% in January, Tesla reaches new lows

Harry Truman

Do I have your attention? See what I’ve done here? I could have done worse: I could have written, without violating any laws, or ethics rules, that “U.S. sales of the Model S, the electric car of high-flying California carmaker Tesla Motors, crashed in January, making a decidedly wrong start into the new year. Tesla currently produces one car, the battery-driven Model S, in a factory deserted by GM and Toyota, and according to estimates by Autodata Corp., the company’s U.S. sales plummeted by 33 percent year-on-year to just 1,000 cars in January. Never in all of 2014 did the company sell so few, Autodata says.” Disingenuous? Yes. False? No, check yourself. Outraged nonetheless? You have every right to be. Truth coming up in a minute.

According to Autodata, sales of the Model S indeed were 33% down in Jan, but an  honest person should never write about this. While it may be technically true, it also is utter bunk. As written a few weeks ago, all monthly Model S “sales” data for the U.S. are completely, and not even scientifically, estimated. The estimates are all over the place, and they differ widely. Estimates don’t even have entertainment value, and they should never be used to draw conclusions, or to write headlines

Why should we care? At an alarming rate, these estimates are used to produce “news” that utterly mislead us.

“Tesla Motor Model S Outsells Nissan Leaf For Second Consecutive Month,” headlines the financial portal Bidnessetc.com. Now there is a completely new perspective of that horrible January. “Tesla Motors Inc’s (NASDAQ:TSLA) electric vehicle (EV) sales surpassed the sales of Nissan Motor Co Ltd’s (ADR) (OTCMKTS:NSANY) mass marketed EV Leaf for the second consecutive month, according to estimates provided by InsideEV,” the site writes. Technically also true. But likewise also utter bunk, which no honest person should put on paper.

What happened? For January, never a good month for EV sales, after a, for tax reasons, always quite busy December, Nissan reported 1,070 sold Leafs, which , as usual, put the car on top of all contenders that sold a few hundred. Then, Insideevs guessed that Tesla could have reported 1,100 Model S as sold in January, if Tesla would report monthly sales numbers, which Tesla firmly refuses to do. Insideevs deftly guessed that exactly 30 Model S more were sold than Leafs, the latter officially. And guess what: We have a winner! Mixing guesses with hard numbers is questionable. Turning them into a ranking already tests the boundaries of honesty. Twisting garbage into a sensational headline is beyond the pale. No CYA footnotes, or throwaway “if trues” make it less  reprehensible.

It’s the second time this has happened. Last month Marketwatch, this month Bidnessetc. Both are no fanzines written by out-of work bloggers. This month’s BS (as in bad story) is brought to us by bidnessetc.com, a business site that prides itself to deliver “the best financial research in the simplest and most creative ways possible.” It is staffed by “a team of more than 100 financial wizards, musicians, rock stars, artists, singers, aspiring philosophers and tech geeks,” under the leadership of Nadir Khan, who says he’s the cofounder of Timescape Global, “a long-short equity hedge fund.” According to the New York Times, that fund closed its door in 2010 after they lost their shorts with a short that went the wrong way.

Who's on first

Who’s on first

Let’s try to find out what really happened to the Model S sales. In this chart, you see monthly Model S “sales,” as guessed by Autodata (blue), Automotive News (grey), and Insideevs (green). The red line represents actual U.S. Model S registrations, as reported by Polk Automotive. Polk is one of the world’s best data sources when it comes to cars. It is owned by IHS, the company that also owns Jane’s Defence, employed by all governments that can’t afford their own CIA. (Which then uses Jane’s.) Polk doesn’t guess its data. It gets them from the motor vehicle departments around the country. Your tax dollars are at work here, so it takes a while. November is the last month for which registrations are available. While you are waiting for Tesla’s official announcements, I can tell you that according to Polk, 12,327 Tesla cars were registered in the U.S. January through November 2014.

Looking at the chart, you can see that all guesstimators are way too optimistic. March through November 2014, the period for which numbers are available from all, 11,202 Model S were registered, Polk says. Generous Automotive News guessed 16,757, and was 5,555 too high. Guessing 14,050, Autodata was closer, but still way off. Kudos to Insideevs: The website outguessed the pros. With 11,600 estimated Model S sold March through November, Insideevs still aims a bit high, but only by 398. Analyzing the data on a cumulative basis is important. One look at the chart tells you that monthly comparisons are meaningless. Reported sales and actual registrations should never be compared month-by-month, there always is an overhang. What also should not be done is mixing apples with lemons, and to announce sales records based on the whims of a harried guesstimator.

Due to serial abuse of the estimates, the DailyKanban calls on Autodata, Automotive News, Insideevs, and all others to stop publishing these estimates. If Tesla does not want to publish monthly data, Tesla’s wishes should be respected.

Absent of official sales data, registrations are the only meaningful data. Those data are available. If I can get them, anyone else can. In an attempt to re-establish honesty, the DailyKanban will continue to perform spot checks.