Ph.D. in NPD

Yesterday, I listened to Paul Thurrott’s podcast with Leo Laporte, and they discussed NPD data. In addition, a number of blog comments have asked questions about it. I thought it might be useful to walk through what NPD is so folks can evaluate the numbers they hear about.
NPD is a huge analyst firm that owns a specialty business tracking consumer Point of Sale data.
Their goal is to get all of the major retailers in the US to share their sales figures each month. So when you hear sales numbers from NPD, they’re actually the sales figures provided to NPD by the actual retailers – not some consumer survey, not a sample, but the real thing. I’ve spent a lot of time swimming in NPD data for the last few years, and in my experience where I’ve been able to compare sales figures from npd per retailer with actual sales figures, they turn out to be scarily accurate.

The nut with NPD is that they haven’t been able to get all retailers to share data. In the mp3 space, they’ve been very successful – the key breakthrough came a year or so ago when they convinced Apple to share both apple store and apple.com sales figures. So how good is NPD coverage? Actually, it’s excellent. They cover Apple, Best Buy, Target, etc… covering 80% or more of all mp3 players and accessories sold in the United States.

They lack a few key retailers, but the only ones that really matter are Costco and Wal-mart. Together, those are around 15% - 20% of this category. They also miss Amazon (but excepting aapl, online sales are also only a small % of the sales). Everything else they miss put together is noise.

NPD data is very high quality, and it’s also amazingly granular. An npd report is a huge list of every mp3 player SKU and model, along with the monthly sales figures going back 2 yrs or more. Once you’ve got that, you can get a zillion useful analyses. What color ipods sold best last month? How much of a sales bump did apple get when they dropped price? What’s Zune’s market share of the mp3 player space? Of the HDD space? How about Sandisk’s? What % of sandisk’s sales are above a $100 selling price? It’s amazingly useful.

NPD also covers a lot of other categories, including networking gear (including networked music/video devices) and mp3 accessories. Those are obviously useful and interesting too, though less sexy.

What can't NPD do? Several key things:

• They don't share per retailer numbers, so you can't find out how many pink ipod shuffles Target sold in February
• they don't cover international, US only.
• They don't cover 100% of retail, so you must gross up based on assumptions to get to the absolute number in the US.

But overall, NPD is a crucial tool in thinking about the space...as long as you keep the caveats in mind.

What's Dave been up to?

Several folks have commented, noting my lack of postings. Others have posted asking me for a view of a "day in the life" of zune right now, and pictures of my workspace.

The answers to these are all related....

We're insanely busy right now. Lots of key decisions being made as we speak. Plus, I've inherited a "team" on Zune that didn't have anyone on it so I've been simultaniously doing what I can of three jobs and hiring into two open heads.

Those decisions are, of course, secret...so I can't really share a day in the life beyond something like this agenda for today:

7:00-9:00 Home taking care of 4 year old with the flu
9:30-10:00 secret meeting
10-11 highly confidential meeting
11:30-1:00 pm out to lunch with finance team, can't say what we talked about
1:00-2:00 Meeting to discuss something I can't mention
2:00-4:00 Meet external vendor who's doing work that I can't discuss
4:00-5:30 Meet with team I can't talk about to plan an offering I can't mention

Not very useful.

There are some other more discussable things...

• NPD data came out - did initial analysis of what's up in the mp3 player, accessories, and AppleTV markets.
• Finished GRAW2 on my xbox - good fun
• Started Crackdown
• Finalizing Car purchase choice...ordering it
• Caring for pitifully sick kids
• Trying vainly to get back into the spring bicycling workout routine

So, that's the news update. I'll post some thoughts on recent news in the industry sometime soon...

And now for something completely different...

I've been shopping for cars this winter; getting ready for the every-eight-years replacement of cars. I really love to drive, but have tended to be value oriented in the past. I've been really struck by what a fantastic array of cars there are on the market for people who love to drive. Some examples:

The MazdaSpeed3 - Amazing little car for under $25k. Nice practical formfactor and 265HP w/ 280 lb-ft of torque. Amazing value - if I were looking to simply replace my 1999 Honda Civic SI, this would be a great contender. I'm actually looking to move upmarket and get a bit more refinement/AWD, so probably not.
Subaru Imprezas. Drove both the WRX and the (amazing) WRX STI. Considering the WRX costs under $25k, it's a great alternative to the Mazda, and includes AWD. It's got a nice beefy engine and has great handling. The STI is simply breathtaking for under $35k - 300bhp and handles beautifully. I ended up turning away from these as they felt a little too barebones for my taste, and a call to my insurance company confirmed my worst fears - these babies are very expensive to insure - more than any other car I'll discuss.

For completeness, I looked upmarket. My first choice was to use my head and get a really reliable Japanese car. I drove two:

Lexus IS350 Sedan ($36k). Wow, was this a nice car. Really well built, beautiful in every detail, high tech but classy interior. I really wanted to like it. Got in and drove it and immediately noted that it's (a) incredibly powerful and (b) not for me. First, the 350 doesn't come with a standard transmission (Really? A performance SKU with no std trans?). They include paddle shifters, but I felt these really distanced me from the driving experience - not a good replacement for clutch and stick. Second, the handling and ride, while superb, was also very distanced from the road. I'd discribe it as driving a really nice paper airplane with a jet engine. Smooth and fast, but without a lot of soul. For me, this car felt like a Mercedes competitor, not really a driver's car.
Infiniti G35s Sedan ($34k). Thankfully, the infiniti came with a standard transmission. This car was also really nice, and handled really well. Great brakes. Awesome price. Felt much better to drive than the Lexus, too. The technology in this car (bluetooth, nav, etc...) is really the best of any car I've driven. The pricing is better than it sounds, since you get nearly everything you'd want in the base price. This one has a pretty good chance with me.

I then went on to try out cars from their Axis partners, the Germans. A note on the prices - the German cars need to get loaded up with another $4k or more to equal the skus mentioned above, so pricing comparisons are tricky.

Audi. I considered the s4 and rs4, but recoiled in horror at the price. I ended up driving the A4 models. the A4 2.0T Quattro ($30k+) lost me right away. After driving the 300BHP Japanese cars, it simply didn't feel spunky enough. I'll talk about A4 handling under the 3.2 below since they were identical in this dimension.
The A4 3.2 Quattro ($36k+), on the other hand, convinced me immediately that the Germans know what they are doing. The power and torque were great, and the driving feel was vastly nicer than even the Infiniti. Until I drove this car, I'd begun to think nobody in the near-luxury class would be able to offer me the "spunky and fun" handling that my Honda Civic SI (and the Subarus) had, but this car was just as fun and more powerful/refined. I liked the interior feel, and the options were nice. This is a great car. Compared to the Japanese, I worry about reliability, but maybe that's missing the point in this class.

Of course, you're silly to look into this category without at least considering the BMWs.

BMW 328xi ($34k). This was my first time behind the wheel of a BMW. It's hard to describe how much fun this car is. Really. It's got lots of pep (230bhp), and handles well. Pretty much all the cars I'm describing do that. But the feel of the BMW is just a bit better than the Audi (which itself is fantastic) and vastly better than the Japanese cars. I realize this whole "feel" thing is very subjective, but for me, the BMW is hard to beat. I also liked the interior of the BMW - not all gadgety like the Infiniti, but with all the nice goodies.
BMW 335xi ($41k). Drove this one (actually a 335i since the xi isn't out yet) for completeness. Almost identical to the 328, but with 300bhp. Amazing Turbo engine gives you massive accelleration even at highway speeds, but with good gas mileage and no turbo lag. Wow, this is nice. The price isn't as much higher than the 328xi as it seems; by the time you price it out with options, you end up just $4000 higher than the 328. Good deal considering the power difference. But in absolute terms, you end up around $50k loaded for this one - a vast difference from the $25k cars we started with.

So, I'm not buying until late summer, but the variety of cars - and the quality and power - that's available at the sub $30k range is really stunning. I haven't driven the VW's and the Hondas, but I'd guess those are great options too.

What do readers think? Have you driven any of these? Longer term experiences? Did I forget anything that I just have to try?

other content fees?

Stillanonymous points out in a discussion thread that many of the mpeg standards - mpeg-4 part 2, AVC/h.264, etc... ALSO have content distribution fees. He's right. I suspect that's a big part of why almost nobody is using those codecs for internet streaming even though they're admirably suited for it. Instead, almost everyone that has shifted from RealVideo has gone to WMV or (increasingly) Flash video. There are some counterexamples, but Flash especially is taking the world by storm right now.

future ala carte services - mp3, wma, or aac?

ok, you've read the heated debate. You've seen the piles of discussion. Now it's time to vote. What do readers want services like Rhapsody, Napster, etc... to use for their "drm free" services that are surely coming? mp3? AAC? WMA? Ogg? Something else? Leave a comment and let me know.

Ignore the many discussions around licensing terms, etc...that's the services' problem. What format do you want to buy music in?

Why are content distribution fees important?

In the last post, I mentioned content distribution fees for mp3 as being important. Here's why I think so.

Traditionally, codecs like mpeg-2 (video in DVDs) charged money in a number of ways:

• if you built a dvd player, you payed a fee
• if you built an authoring tool that created dvds, you payed a fee
• if you printed a disk, you payed a fee.

These were all fine, because all competitors in the space payed the same amount, and thus it just went into the cost of your DVD player, or copy of Terminator 2, or whatever. Content producers recouped by raising prices, and everyone had to raise it by the same amount because there's no alternative to DVD.

Streaming is different from DVDs. For example, mpeg-4 part 2 was promoted as replacing Windows Media and RealVideo for web streaming. But the vast majority of web streaming was free of charge (enabled because WM and RNWK didn't charge a fee for distribution). But if a site switched to MPEG-4 part 2, there was no way for a website to recoup the per-stream costs. Imagine if Youtube owed a few cents per stream to macromedia for distributing content on flash? And cnet, cnn, etc...? They'd switch codecs pretty rapidly, or else drop free streaming entirely.

Music services are also different from DVDs. The margins are razor thin - basically nonexistent - so if Apple uses AAC (which has no content distribution fee) and Napster uses mp3 (which does), Napster is at a 2% disadvangage vs. Apple, which could be the entire margin on the business.

It's interesting how choices made in the licensing regime for a technology can make or break it's adoption for different applications.

mp3 content fee!

The many posts below have taught me two things:

• The surprising fact that there are AAC "fanatics". For a codec? Is there an mpeg-1 fan club? an h.263 shrine? ;) This probably has more to do with Apple fandom than AAC itself. I mean, AAC is a nice audio codec and all, but...
• The really important thing is that mp3 has a content distribution fee. The Thompson site says you pay 2% of revenues from a service - or $0.02/song - to use mp3 in a service. The last time I saw a content distribution fee, it was on mpeg-4 part 2, and it pretty much killed that standard. Thankfully, in h.264 mpeg dropped that requirement for free distribution.

Surprising, and important.

wma licensing

I didn't orginally post, but had to look this up. Looks like wma costs:

• $0.10 per unit, up to a maximum of $400,000

This per http://www.microsoft.com/windows/windowsmedia/licensing/final.aspx#WindowsMediaAudioStandardVersion_Final

licensing bodies and costs

A reader (thanks, Randy) points out that licensing is complex.

• AAC isn't licenced by MPEGLA, it's licenced by Via Licensing (aka Dolby).
• MP3 is licensable from Thompson, not MPEG.

Just trivia, but interesting. Pricing isn't necessarily so trivial. at “published” rates:

• For portable devices MP3 is $.75 per unit.
• For AAC, if your annual volume is less than 400,000 per year (100K per quarter), $1.00 per unit. If your unit volume is between 400,000 and 2 Mil units per year (200,000 to 500,000 per quarter), it is about the same $.74/unit. Above two Million units a year, AAC gets cheaper.

This is one of they key barriers to hardware adopting new codec technologies - each additional codec eats into precious hardware profit margins. But if something is ubiquitous and necessary, everyone pays it and prices pop up to make up the difference.

More than you ever wanted to know about audio formats

Thought I'd lay out some fundamental facts about a few audio formats - some comments show some common misperceptions.

• Mp3 – the jpeg of digital music. Supported by everyone, everywhere. Mediocre compression efficiency, but utterly dominant in personal libraries because users can simply rip to a higher bitrate. Licensable by anyone for a low fee. Universal compatibility is mp3’s key feature.
• AAC – designed for MPEG, by MPEG. Largely associated with Apple at this point, AAC was designed by a licensing body, complete with patent pool, etc.. Licensable by anyone for a (relatively pricey) fee from MPEG-LA. High compression efficiency. Compatibility is more limited, but the iPod brings AAC support to 10s of millions of iPods (and PCs). iPod ubiquity is driving improved compatibility over time.
• WMA – Microsoft’s audio format. Adopted by most mp3 player skus (but not by most mp3 player units, thanks to the iPod..). Freely licensable for a low fee. Also associated with WMDRM, which is supported by playsforsure devices. High compression efficiency, good compatibility story, but not as good as mp3’s.
• Ogg, FLAC, etc… - Open source codecs. Generally good compression efficiency, but very low support from services and devices. To use these is to enter untested legal waters, but they’re very interesting, and nobody is asking for money for them...yet.

I’m sure these will generate some fun discussion, but thought it worth laying them out for folks to read.

New policy - Now with deleted abusive comments!

I was reading some of the comment strings on the last post, and got cranky about some of the abusive language from anonymous posters. I've strictly refrained from deleting comments, but some of these were over the top, plus from the cowardly veil of anonymity. It's too bad, because some of them had interesting points of view embedded in the abuse, and I'm always happy to learn from comments. However....

In the interest of transparancy, I'll announce that I'm going to start nuking anonymous posts that I consider to be over-the-top nasty.

"Clear" AAC tunes

In light of the itunes/EMI deal, lots of folks are asking if Zune supports clear AAC tracks. The answer is yes - today's Zune software and device both already support AAC tracks, including those that iTunes will presumably be selling.

I think it's interesting they chose to do AAC instead of mp3 files. I suspect Apple's systems were probably just not set up to "sell" mp3 files - for now.