Archive for the ‘Business’ Category

making sense of the mobile web statistics

Saturday, May 1st, 2010

It's easy to decide to ignore all the statistics out there, or to just go with the one you like best and ignore all the others. But if you analyze the sources a bit more closely, you will get even better insight than if all the sources agree.

The Contenders

There are two major categories of statistics: one in which data is automatically collected across some subset of the Internet, the other in which data is collected by asking some type of person.

Automatic collection

Ideally every request on the web, whether for a whole page or for an object of some type, has its data stored. What device, what browser, what operating system, what country, etc. Alas, that data doesn't exist. So all automatic systems are collecting only some subset.

There are two major types of automatic collection:

  • Analytics installed on individual web sites using a specific platform, then aggregated into metrics. This includes normal analytics, like StatCounter. It also includes advertising platforms, most especially Millenial Media and AdMob.
  • On-device reporting, such as that from Nielsen.

Most (but not all) of the analytics providers have solved the collection problem of devices not using Javascript. If you are only concerned about "smartphones" then most of the browsers these days have Javascript.

For any automatic system, you have to consider how the data is gathered. The advertising platforms (Millennial, Admob, etc.) are only collecting data from their network. So their numbers are affected by:

  • Global coverage - each covers some countries better than others
  • Ad formats - Millennial has special ad formats for iPhone, and does not promote Android to its advertisers. Admob, on the other hand, has special ad formats for Android. If you were an advertiser looking to target both platforms, which would you choose?
  • Advertiser type - Millennial seems to cater to bigger brands and has less of a self-service long-tail client base than does Admob.

More fun: The Weather Channel and possibly other large mobile-savvy companies do their own ad sales, and aren't represented anywhere above. And some sites will have ads on every page, where others will limit the ads. The former will have greater representation in the data.

Interviews

Nielsen, Pew, and some others use more traditional interview and survey techniques. By and large, these are done by companies who are focused on getting statistically valid data. Of course, as with any research, you have to investigate what they are asking and who they are asking it of. In particular, people don't know what type of phone they have, so asking them will result in errors.

At least one company asks mobile operator store managers which devices were sold in the past week or month. This is interesting data but doesn't tell us anything about how the devices are used.

I am especially fond of the Pew research, because it focuses on behavior, demographics, and actual use. It's helped our clients understand that many of their demographic preconceptions are wrong, and that their teenager preconceptions are merely incomplete. But that doesn't tell us what devices are being used.

Which Browser Has the Highest Mobile Web Use?

Millennial says iPhone. Admob now says Android, in the US. And StatCounter has said Opera for all but three months in 2009! Well, until you notice that iPod Touch is measured separately from iPhone.

One more bit of complexity: both Admob and Millennial serve ads to apps and web. And Admob serves ads to the top 10 free iPhone apps, so you would think there might be some bias there.

Source: StatCounter Global Stats - Mobile Browser Market Share

That's the wrong question for most of us. We instead want to know what smartphone/browser/phone has the highest use, in our audience.

What you should to is consider your market, and put together the data accordingly. I expect that StatCounter has disproportionate representation in India, where Opera and Nokia represent something close to 95% of the browser visits. (Admob largely agrees) Are your visitors in India?

In Indonesia, another high-volume mobile web market, there are a lot more Sony Ericsson devices (20% per Admob) but they don't make a blip on the StatCounter site. Apparently all the Sony Ericsson users are using Opera. Are your visitors in Indonesia?

StatCounter shows an uptick in Opera and a larger uptick in Symbian in the past two months. Is that a shift in adoption, or a shift in mobile web moving more into Asia? Admob data strongly suggests the latter. Or maybe Admob's sales in the region have gotten better.

Are you a premium brand? My hunch is that you should trust Millennial's data more, that their ad behavior better matches your customers' behavior.

Are you a long-tail offering? Then combine StatCounter and Admob.

Are you planning a free app? For now, I'd go with Admob. That'll change as iAd comes online, of course.

If you really want to know highest use for purchasing purposes, trust Nielsen the most. They don't measure every site out there, so do a bit of research before making ad purchasing decisions. Check out this mobile marketer article for more on choosing an ad network based on reach.

If you are deciding where to focus your development efforts and want to use just one source, again choose Nielsen. But I hope you'll look a bit deeper based on what we talked about here.

history is context

Friday, April 16th, 2010

Despite historically being a personal computer maker, Apple is clearly moving to be some sort of “digital lifestyle” company. Forget the cute name change (dropping “Computer”) and other direct comments about this. They are all self-serving and much more can be learned from examining behaviors.

A key “behavior” of a company is the products it releases. Let’s look at the highlights of these “non-computer” products:

A (mostly) ARM based platform, centered around pen/touch interaction with specific appeal to the PDA, mobile communications, networking and graphics. Add ons and applications are supported with a custom SDK, allowed them to be written relatively easily in an OO language.

Devices that support the platform include simple tablet-like PDA/entertainment devices, smartphones, ruggedized devices for verticals and larger devices with integrated keyboards.

Any of those specs throw you? A keyboard on the larger device perhaps? That’s because this is not iTouch/iPhone/iPad, but the Newton platform, and refers to the MessagePads, Seahorse, Tarpon, Schlumberger Health Terminal and eMate among the many devices made to support the platform.

And this all started in the 1980s.

I periodically rant in Twitter about the stupidity of tech pundits. Such that I often don’t read, watch or listen that much, as they infuriate me. This is exactly why. No one [well-read, that I know of] has talked about the above history in any but the slightest detail basically since it was happening.

But a straight line can be drawn from Apple’s computer platforms, though the Newtons to the iPhone platforms of right now. And if you act all pundity and think about what they did, that’s called analysis. You can come up with some useful intelligence from it.

My key takeaway from Apple is actually a public statement, the Digital Hub concept. Delivered way back in 2001, and which I think they are clearly still working from.

You get one desktop computer, you attach PDA-type devices, entertainment devices and portable devices to it. You attach to the internet. And you get to continue being productive (work) as well as being entertained. Remember, this was the Information Superhighway era and there were Info Appliances (all to often, actual appliances like refrigerators with computers embedded) proposed and sold that would do specific tasks and network in their own way. It was not clear that the general purpose computer, the PC had a future outside of offices and servers. In fact I am a failed futurist – I thought the PC was dead and we’d all have info appliances for specialized tasks, and iPad like things and smart refrigerators. Instead, we have a faster, shinier ten-years-ago sort of computer world.

Anyway, back to Apple. They are still working on this whole philosophy and it’s easily visible in their mobile platform. Who else would require tethering to a computer to get software updates? Who else would assume (not just require, but assume) that the center of your media library is the computer, not the portable device. Etc. The core information architecture follows precisely from this assumption.

And – without getting too deep into it – their strategy for OS and application control has been unchanged since at least the Mac Clone business was killed in 1997. This extended straight to the portable platforms (nothing at all on the iPods, and strict control only when the market clamored for, and hacked into, ANY openness on the iPhone platform). Even before this, the little Pippin software was all routed through Apple. It even informs Apple’s own apps, like the iWorks suite; pretty, but built from a singular point of view. Trying to move outside the intended scope is difficult or impossible.

It’s not a shock. It’s not a regrettable surprise. And no matter how much I like to argue they could add something like an “untrusted app” feature, or better type controls, or a color picker that isn’t insulting to colors, without sullying their brand, it’s really not likely. Because this is now baked into the DNA of the company.

We spend a lot of time developing this knowledge at Little Springs. We have a wall of old phones and spend a lot of time showing what even 2-3 year old devices worked like to the sometimes annoyingly young staff members and interns. Some of us have been in mobile specifically for over a decade, and that’s just forever. Just myself, I’ve been on projects for the first musicphone “ever” and the first camera phones in the US, and built an app store something like 10 years ago.

All of this we try to keep in mind, and pass on to the rest of the company, and when we get our act together, mention in the blogging and on the wiki so everyone in the world can share in it. The next time you are trying to figure out something, ask one of the old timers around your office (or, email us I guess) to cast their mind back, and see what you can learn from the not-really-very-old history of technology.

And it doesn’t just make you feel good. Knowing what has been done in the past helps you exploit the concepts, or avoid those mistakes if they are fundamental (e.g. not just that the technology wasn’t advanced enough yet). Knowing what the competition is doing lets you predict their next move, and understand their current moves. For example, I’d say you are totally safe assuming Apple will continue to place the Mac at the core of their strategy, and protect it at significant cost. If you can exploit that, you have a leg up; you can design and market for the appropriate niche.

This works for any product – I just used Apple as a well-gossiped case. What are you competing with, or wondering about, or building that you think is all new? Are you sure it’s totally fresh and new? Are you sure you understand the competition as well as you can?

video on the second, third and fourth screen

Wednesday, March 31st, 2010

I periodically hear some director or pundit talk about how terrible it is to watch a movie on a laptop or mobile phone. And at face value, it makes sense:

Sizes of devices absolutely compared.

That phone is so tiny! Why would anyone watch a movie on that? And the laptop is pretty marginal. Until you actually take into account reality, and users and math.

Sizes of devices as viewed.

People automatically adjust for what we (possibly incorrectly) call angular resolution. Entirely aside from phones being hand-held, they get moved to the right distance to see things correctly. Broadly speaking, video display ends up occupying the same field of view regardless of the device size.

This is why people actually do watch video on phones, and will continue to do so.

micropayments and so-called micropayments

Wednesday, February 24th, 2010

Somewhere in my Twitter feed, I was pointed to Rita McGrath’s Why I Hate Micropayments on Harvard Business Review. A nice read, and points to several problems. But I’m not ready to throw the baby out with the bathwater.

When originally conceived, micropayments were pennies, fractions of pennies, or maybe small fractions of dollars. This has unfortunately migrated into $0.99 or even $2.99 “micropayments.” Maybe these are “centipayments?” Nah, too geeky. “Minipayments” is better.

I think the minipayments evolved because you can’t make micropayments work financially for a single site. With credit card fees starting at $0.30, a $0.05 payment isn’t going to fly. Even Paypal’s current offering is $0.05 + 5% for micropayments.

I’ve seen some attempts to create cross-site micropayment systems, but they were way too small and folded soon after. At one point I had sent $10 on each of three systems, but only ever got $3 worth of content out of any of them.

Premium SMS is great, except that you only get at best 25 cents on the dollar.

Problems with minipayments

Right now, minipayments have business and user experience problems.

  • They are large enough to make you think about clicking, to actually track the money. I could have gotten a song for that! Or a cup of coffee! Or a turnpike toll!
  • Each micro-purchase is made with a separate financial decision, unlike turning on a light switch, sending a text message, or driving to the store. Each of these other decisions have financial consequences, but few of us think about it.
  • Most of these decisions also have a high time and attention cost. Entering credit card data, for example, is slow. Even password entry creates a barrier to entry. These extra user steps remind the user of the dollar cost as well.

There are places where minipayments are making sense. Apple’s iTunes store is clearly an example. Arguably, other app stores are also successful.

These success stories have addressed part of the minipayment problems. Customers have a single iTunes account, they aren’t really paying attention to how much they are spending, and Apple is left to get the money to the content creators.

But a central store for web content simply won’t work. The “store” needs to be highly distributed.

Other successful models

For sites with known value, small annual payments are useful. Flickr’s professional account is $25 for a whole year; $2/month is invisible to nearly all.

Monthly fees are more problematic, as customers get reminded every month of the cost. We subscribe to a few services, and seeing $50 here, $25 there, and $100 over there every month is painful. I find myself stripping down services like this.

Annual payments could work for some content companies, but customers would have to know the value. I’d happily pay $20/year for Harvard Business Review content, but I will not pay $4/article on nearly any site.

Making micropayments work

In short, we need an ecosystem for acquisition and payment of content, especially long-tail content of all flavors. If done correctly, the ecosystem could allow bands to sell their songs directly, application developers to sell their apps directly, and of course some web pages to be behind paywalls.

Micropayments need to be beneficial to the content owners, the payment providers, and to the end users. Here’s one way how:

  1. A large brand, such as Visa, Paypal, Google, or mobile operators, creates a micropayment platform. Actually, we need to see several of these.
    • Mobile operators would simply use their payment gateway. For more, check out Visionmobile’s Why mobile can bring back the value to the internet
    • Google and Paypal may need to require the account to be primed with a small amount of cash, which can be withdrawn at any time and also applied to other purchases.
    • Visa and Mastercard could make micropayments available to good customers carrying a balance, or perhaps add an annual $10 fee
  2. Payment providers create a signup for content owners intended to be fast and simple, more like signing up for Adsense than for an App Store.
    • Protecting against fraud may involve a refundable deposit, scaled against projected sales volume, for unknown providers
    • Content owners should sign up for as many payment systems as possible
    • Payment provider provides buttons for their call to action
  3. Credit or debit cards associated with micropayments get charged once per month, or similar, perhaps just refreshing the account balance. It could vary by provider.
  4. A pre-pay option could be cash-based, much like buying minutes
  5. Content behind a paywall would have an array of payment buttons next to it, much like the social media share buttons on blog entries right now.
  6. Purchases are one-click or click + passphrase, based on security settings
    • Purchases are limited to one per 3 minutes (fraud protection)
    • More than, say, $5 spent in an hour triggers an authentication request
    • No purchase can exceed $3 without an authentication request
  7. The payment provider returns an approval code that allows the content to be displayed.
  8. Payment providers have easy-to-integrate code for hiding and revealing the content, ideally using a standardized micropayment API

By batching the transactions to a weekly or monthly level, we reduce the effects of transaction fees. The same function makes the price effects of the purchase less visible to the customer.

By allowing purchases with a click or very short code, we reduce transaction friction for end customers. By providing several types of accounts, we provide flexibility for end customers to choose the right type, such as mobile phone, debit, or prepaid.

I agree, this isn’t the way to save the newspaper industry. But it’s more likely to help than to require that I pay $5 for a week’s access and have to sign up for a specific site. And it could be highly disruptive for a number of long-tail publishers.

Paper or Plastic? e-readers vs mobiles vs book

Wednesday, February 10th, 2010

With all the latest announcements for e-readers and book pricing, I’ve been thinking a lot about the nature of reading recently. It seems like each company’s reading experience presumes everybody and every bit of content is the same experience.

Amazon is quite proud of their $9.99 pricing for e-books. They claim that it is cheaper than paper. And it is … for one type of reading. If you read mass market hardbacks or trade paperbacks (the ones that are the same size as the hardbacks) then $9.99 is cheaper than $15.99 or $25.99.

I do that type of reading, but only for professional-related books. These books I might want to learn from, cross-reference, make notes for adoption at the office, and so forth. I will read while traveling, but not relaxing in the tub. I don’t really share these books with my friends, though I share some of the ideas with my colleagues.

In contrast, my personal reading time is spent almost exclusively with cheap paperbacks. They’ve increased in price to $7.99, which I think is too much, but I pay it anyhow. I actively share these with people I live with, and share with friends. These are the books I take into the tub, and are more likely to come to bed with me. I keep them for a long time, but if they get lost on a trip or wet in the tub, it’s okay.

Several of these books currently live on the shelves of the people I lent them to, and while I’d like to have them back it’s completely livable. You can tell which series I really enjoy, because I no longer have the first book in the series. It’s been loaned elsewhere.

Then there are newspapers and magazines. This content is temporary by its nature, and at least for me is more akin to skimming online news and blogs.

These are not the only types of reading out there. One of the key lessons in graduate school was how to read research papers. Reading for learning or book clubs might also look very different. Reading business documents, at least in my product development world, really needs a large screen and a strong social connection. The documents do not live in a vacuum, and their business context must be considered while reviewing the product requirements document. I suspect that legal reading is different again.

Each manufacturer/provider has presumed one type of reading. Kindle is targeting trade paperback/hardback reading, largely for pleasure. The Que is designed for business use. Hearst’s Skiff is targeted more at newspapers. Apple is going after content deals galore, including textbooks and newspapers. None appear to target my pleasure-reading needs.

I’m not really worried about device fragmentation in the e-reader market. We know how to design and develop for that. I’m more worried about the purpose fragmentation. Is one going to win?

Clearly the mobile phone will continue to win here. The Kindle’s ability to continue reading on the phone where you left off on the device is brilliant. But the problems with sharing and moving content to other places still plague them; I only want to purchase material I know to be temporary. This is an industry DRM problem, but Kindle appears to be extra protective. Still a no-go for me.

In the meantime, I’m still reading the occasional book on my phone, and carrying several paperbacks on trips.

Are you wasting your mobile advertising dollars?

Wednesday, February 3rd, 2010

I was in a game on my Android. I kept seeing this kinda-teasing ad for a game; I thought it might be relevant to me as a casual puzzle game player.

Finally, after hours of play, I clicked on the ad. Many users do this, in higher rates than on the desktop-based web. The advertiser paid for me to click this link.

What should the link do?

  • Take me to a landing page directly related to the ad
  • Already know my region and device
  • Be optimized at least for mobile
  • Have a call to action on the page
  • Give me access to other information in mobile-accessible format

What did it do instead?

  • Site home page, not landing page
  • With no information about what the game was, not even platform(s)
  • Demanded that I go through a two-step process to enter my region
  • Took me to the product information page, coded in Flash