Russian-based search engine Yandex this morning released a new mobile app called "Wonder." It's currently only available for the US market and right now only on iOS. It uses speech recognition from Nuance and social data from Facebook, Instagram, Foursquare and Twitter to provide search results refracted through social networks.
The kinds of queries Wonder envisions are those such as "tech news stories liked by my friends" or "restaurants near me visited by my friends." The app can only be used in landscape mode. It offers a visually polished UI but generally poor search and user experience (it would be better on the iPad). Unless there are some dramatic changes it won't be widely adopted by consumers.
Indeed, Wonder is no substitute for Google or Yelp or Facebook's new local and general search capabilities. What's interesting and significant is that it does illustrate broader adoption of social data as a filter and mechanism to personalize search results. The app is also consistent with the embrace of voice as a primary UI and capability.
Wonder offers a hint of a personality there but it's not a full blown "assistant" like Siri or Speaktoit. However Amazon's acquisition of text-to-speech specialist Ivona is a move to bring a Siri-like assistant feature to Amazon's Kindle tablet devices. (Amazon previously purchased speech provider Yap.)
Amazon already had text-to-speech for Kindle but Ivona offers a smoother, human-sounding voice capability that can be deployed for a range of purposes and use cases. And like Wonder it reflects the degree to which speech has become a critical "must-have" function on mobile devices.
By itself, however, speech is not enough. Increasingly there must also be a "personality" (assistant) to go along with the raw speech-processing capability. This is the impact of Siri on the broader marketplace.
My colleague Dan Miller brings a different perspective to the Ivona acquisition. He sees it in the larger context of speech-industry consolidation.
Update: TechCrunch says that Facebook has blocked Wonder's access to its user data while the companies negotiate about access.
Yesterday comScore reported that Facebook had the number one US mobile app of 2012. It beat out perennial leader Google Maps for the top spot. As an aside, Google Maps now has more usage in mobile than on the PC according to comScore.
There's nothing necessarily surprising or remarkable in Facebook's rise to the top position. It has long been one of the most popular apps and the app with the greatest monthly engagement (time spent).
What may be most significant here is the way in which Facebook's victory over Google operates as a kind of "metaphor" for the differences between PC and mobile usage. Search is not the center of the mobile experience as it is on the PC -- although local search, represented by Google Maps, is a critical and hugely popular function. Of course Facebook Nearby is a local search tool and has significant potential if Facebook invests further and continues to develop it.
Notwithstanding Facebook's win, Google still dominates the top 10 in comScore's chart above.
As widely discussed Facebook's challenge is to fully "monetize" all this mobile traffic/engagement without negatively impacting or "corrupting" the user experience. Search will help the company do that as would a mobile ad network for third party publisher sites that used Facebook user data. (There are some privacy issues and potential challenges there and it's on hold.)
Facebook has "display" ad units for mobile that are helping to quickly ramp its mobile revenues. The company allows mobile-only targeting as well as combined or cross-platform targeting for PC and mobile, thereby simplifying the mobile ad buy for marketers.
Many analysts simply assume that mobile advertising will follow the well-worn path of PC advertising, only perhaps in a more accelerated fashion. Thus you get mobile advertising forecasts that show a relatively smooth progression of ad budgets into mobile, with search and display being the two main ad categories (calls fall into both). There's a longer post to be written about these assumptions and why they may not play out as expected -- especially with respect to location-based ads on mobile devices.
Overall there are relatively few mobile search impressions available outside of Google. So most of the ad inventory being sold today is some form of mobie display. However there's also something a "war on mobile display." That's really about the business model: CPM/CPC vs. CPA.
It's largely being waged by firms whose business models that are not CPM or CPC based. Companies that use a pay-per-[app]-install or other CPA models have attacked CPM or CPC-based mobile display with the idea of the "fat finger problem."
The argument is that a huge volume of mobile display clicks are simply mistaken or perhaps even fraudulent in some cases.
Source: Trademob (9/12); based on analysis of 6 million mobile ad clicks
There's also the idea, often discussed, that consumers don't like mobile display advertising and consider it to be just a notch above spam.
More recently Marchex, which has transformed itself into a call-based advertising platform and network, asserts that the overwhelming majority of mobile display impressions and clicks are nearly worthless. In a study, released in December, involving six major display ad networks Marchex found that it took almost half a million ads to generate one "quality" phone call:
We examined a set of mobile display ad campaigns across the six largest mobile display networks to investigate the real, measurable performance of these ads. The call to action on all advertisements was a phone call. Performance was based on the number of high-quality calls driven by the media investment. Marchex defines high-quality calls as those that do not include misdials and spam; existing customers looking for support services; and unproductive calls (e.g. too short).
Our advertisers included national, branded businesses in Education, Insurance, Home Services and Entertainment. We conducted the study on major mobile ad networks and employed media tactics ranging from highly targeted to broad buys. Ad spend was distributed across networks and advertisers to ensure statistically valid conversion results on the back end.
Marchex said that in its test it took 494K impressions to generate 2,481 clicks, which in turn generated only one "quality" phone call (as defined above). That single call effectively cost $302 according to the company, because of all the wasted impressions.
Source: Marchex (2012)
I exposed these findings to one mobile ad network, which disputed them and said on its network the ratio of impressions to qualified calls was much smaller: 15:1 rather than 494K:1.
The Marchex argument is that it's simply cheaper to buy calls directly than to buy mobile display impressions. The company's study needs to be replicated before we can conclude that Marchex's findings are valid across networks. There's also the argument about awareness vs. direct response -- most national advertisers currently are just seeking broad awareness and scale.
Regardless Marchex's findings and the other data above collectively fuel pervasive doubts about the value of mobile display advertising.
Yesterday afternoon Google announced Q4 2012 earnings. In almost every respect it was a spectacular holiday quarter for the company. Consolidated revenues (which include Motorola) were $14.42 billion, an increase of 36% over 2011.
Google made $50.2 billion for the full year, crossing that revenue threshold for the first time. That compares with $37.9 billion the company made in 2011.
However the average price that avertisers paid Google per click (CPCs) decreased 6 percent vs. Q4 2011. That was a smaller decline than in the past, which could be seen as a positive.
The CPC YoY drop is because more clicks are now coming from mobile devices and advertisers are paying less for those clicks. According to a report released yesterday from marketing firm The Search Agency, CPC prices for paid-search ads appearing on smartphones are well below comparable ads appearing on tablets and PCs (see graphic below).
In Q4 mobile search clicks were worth less than 50% of what marketers paid for PC search clicks according to the data. Why are marketers paying much less for mobile clicks when mobile consumers are often much better prospects and customers than PC users?
There's less competition currently for mobile clicks than there is for PC search clicks. Because Google's ad system is an auction that necessarily affects pricing. But more than that many advertisers are unwilling to pay more for mobile clicks because they don't trust them and/or can't calculate a mobile ROI.
Source: The Search Agency
Many search marketers, especially brands and large advertisers, rely on automated systems that calculate paid-search ROI based on some pre-defined conversion event. Those conversions can be a variety of things but frequently they're e-commerce transactions or, in some cases, phone calls.
PC ROI calculations are generally flawed because they usually don't or can't capture online-influenced offline buying. Accordingly the system and the marketer only see online events but not the far larger collection of offline purchases and activities (e.g., store visits) that are driven by online and paid search advertising. The problem is even more pronouced for mobile, however.
Because there are relatively few mobile commerce transactions -- though there are plenty of phone calls from mobile devices -- marketers simply don't see the "latent" conversions that happen in the real world or later on another screen, such as in the case where someone does research on a mobile device and later buys on a PC or tablet.
As a result of this varied, multi-screen consumer behavior marketers aren't able to correctly perceive or attribute ROI and accordingly value mobile clicks. While this represents a "buying opportunity" for advertisers that know the true value of mobile the majority of advertisers are undervaluing mobile clicks. And that's reflected in the average CPC declines that Google has been reporting.
Yesterday Nokia announced "better than expected" Lumia sales. Overall the company said (in these preliminary results) that it sold just over 86 million mobile devices. Among them were 16 million smartphones, including 4.4 million Lumia handsets. The remainder were legacy Symbian devices and new lower-end Asha devices.
Asha phones are somewhere between a feature phone and a true smartphone. They're designed to be low cost and intended for emerging markets such as India. They would see little or no success in developed markets like North America or Europe. Indeed, they're not directed toward those markets.
In Q2 and Q3 2012 Nokia sold a combined total of 6.9 million Lumia handsets. The troubled-company's stock was up yesterday and this morning, having seemingly beaten a very grim Q4 forecast. And some financial analysts are hailing the results as the beginning of Nokia's long-hoped-for turnaround.
Any celebrations are premature however. According to Kantar Worldpanel Comtech research demand for Windows Phones is uneven and limited.
In the US Windows Phones continue to lose share and have failed to capture consumer interest. The story is somewhat different in Europe, however, in part because of the legacy of Nokia's strong brand. In the five major EU countries Windows saw aggregate growth of 1.7%.
The markets where Windows Phone gains have been meaningful are Italy, Spain and the UK, according to the Kantar data. In Italy, for example, Windows Phones gained almost 8 points and now have an 11.7 percent share of the smartphone market.
While there may continue to be modest growth for Nokia with Windows Phones, it's fairly clear that they are unlikey to power a full recovery. What Nokia really needs to ignite growth is to add Android devices to its lineup.
Remarks earlier this week by Nokia CEO Stephen Elop suggested that the company could be open to using Android:
In the current ecosystem wars we are using Windows Phone as our weapon. But we are always thinking about what's coming next, what will be the role of HTML 5, Android... HTML5 could make the platform itself -- being Android, Windows Phone or any other -- irrelevant in the future, but it's still too soon [to tell]. Today we are committed and satisfied with Microsoft, but anything is possible.
Contractual agreements with Microsoft probably would make Android "diversification" unlikely in the near term unless Windows Phones sales fell below a certain threshold. Given the modest momentum around Microsoft's OS Nokia will probably stick with Windows.
Yet if the company were to offer both Android and Windows devices it would see its fortunes improve more rapidly -- much more rapidly.
Nuance Communications, which provides speech recognition services for enterprises (and increasingly consumers, including the Swype keyboard) released a “mobile assistant” survey in connection with CES. The survey of roughly 1,000 US adults found that 75% of respondents had their mobile devices (presumably smartphones) “always on them” or “at hand.”
Among the 90% of survey respondents that reported they had some sort of assistant capability on their phones (not defined in the survey results), a majority (60%) said they used that assistant daily. The following were the most common use cases:
This survey implies satisfaction is relatively high with these assistants. More than 80% of respondents indicated if they could they’d want the “same mobile personal assistant” with them at all times, across all devices and use cases: phones, tablets, PCs, cars, TVs, apps and so on. Accordingly the survey was partly intended to support Nuances "cross-device persona project" called Wintermute, which the company is showcasing at CES.
Using your unique voice print, the system remembers who you are and "follows you from one device to the next, remembering what you like, what you’ve been doing, and where you’ve been." This is in a way a voice-version of what Google is trying to do in asking people to sign in to the Chrome browser so that it can monitor them across devices. In Google's case it's for the purpose of personalizing search results, serving better ads and collecting data on user behavior. Nuance seems to be focused more directly on improving the user experience.
Interestingly the survey also found that respondents had emotional connections (to varying degrees) to their assistants:
More than half of all respondents cited a personal connection with their mobile personal assistant. Women actually name their mobile personal assistant more than men, with 71% compared to a close 66% of men . . . 73% of men feel comfortable asking their mobile personal assistant for directions but 79% of women ask for help more often.
The materials I received don’t provide detail on whether the assistants in question are Nuance products (i.e., DragonGo) or Apple’s Siri or Google Voice Search/Now. It's not clear how specifically the "personal assistant" idea was defined in the survey instrument.
The concept of the personal virtual assistant has been around for quite a long time, using a range of technologies and approaches. Yet crystallized in the public mind with the advent of Siri. Nuance, which provides speech recognition for Siri, recently introduced Nina -- a white label Siri-like assistant for enterprise customer service applications.
My colleague Dan Miller recently issued an expansive new report on personal virtual assistants and their adoption in the enterprise and on consumer devices.
Update: Nuance has reportedly acquired VirtuOZ, which is a provider of virtual-assistant enterprise customer care solutions with a PC focus. The VirtuOZ online assistant will be enhanced and improved by Nuance's speech recognition capabilities and Nuance's Nina offering will help expand its reach into online customer care.
Mobile handsets already outnumber PCs across the globe. However, in terms of internet access (including apps) PCs could fall into third place after smartphones and tablets. That would represent a radical change from today, where PCs still represent the most common way that consumers in most developed countries access the internet.
Hardware monitor NPD Group is predicting that this year tablet shipments globablly will exceed notebooks (laptops). Here's the projection:
NPD also takes a stab at forecasting penetration of tablets by screen size. However this is somewhat less important than the fact of tablet adoption.
Rather than the cornucopia of screens displayed in the graphic above, we're likely to see standardization around three primary tablet screen sizes:
The general publisher response to the rise of tablets is probably going to be responsive web design, which has limitations. However tablets will further reinforce app usage among consumers, though surveys and behavioral studies show more mobile web usage on tablets vs. smartphones.
Tablet adoption also ends the reign of Microsoft as the most important company in the PC universe. Windows 8 and Surface adoption to date has been quite tepid. And the inevitable availability of Office for tablets removes yet another incentive to buy a PC.
The "big takeaway" here is the simple fact of more tablets than new PCs. PCs will remain prominent in the workplace and among business users. However for consumers they will see less and less "face time." Tablets and smartphones will become primary internet access points in the home, with PCs being used for more limited and specific things.
One of the challenging things for marketers these days is to figure out how to efficiently reach consumers on the growing array of screens they interact with. The growing complexity of consumer behavior and the interplay among devices is dizzying.
Last year Google did some terrific research about the parallel and sequential usage of smartphones, tablets and conventional PCs along the path to purchase. The company found that 90% of US adults surveyed used multiple screens during the day. It's really challenging to track this behavior in real time let alone create coherent, integrated campaigns to address it.
One of the central behaviors identified in the Google research was the multi-screen path to purchase. Consumers often start on one screen but complete transactions on another. The behavior wasn't random, however. Smartphones were found to be the most commonly used screen but people chose different screens depending on the context and nature of the task at hand.
Harris Interactive has released similar research that reflects different user preferences and behaviors depending on the particular screen and use case. Harris found overlapping usage scenarios but also consumer preferences for one screen vs. another in several instances.
The survey sample consisted of 2,383 adults, about 42% of which owned a smartphone. However that's lower than the US mobile average of 50%+. The data were collected in November 2012.
The question fielded was: "Thinking generally about your media and communication behavior on a smartphone versus on a computer, please indicate which of these actions you regularly perform on each." Multiple responses were permitted.
In some cases smartphones tended to be used more and in others PCs dominated. Unfortunately Harris didn't ask about tablets.
General activities (penetration/usage):
Social media usage (penetration/usage):
The presence of children in the home was correlated with increased smartphone activity across almost all categories of activity.
In looking at these data one can see that certain kinds of activities, better suited to smartphones (texting, map usage, checking in), are more often performed on mobile devices. However activities that require larger screens or where the mobile user experience is sub-optimal, favor the PC (e.g., product research, purchasing).
Independent analyst Ben Evans has teased out a range of Facebook mobile usage and user data, partly derived from the company's own public statements and partly from his own calculations. You can read what he says here. Below I use some of his data and one of his charts.
Mobile users as of Q3 2012 (mostly public numbers):
Accordingly, roughly 44% of Facebook's global user base doesn't access the site on mobile according to the company's own data. However that figure is likely to get smaller over the next 12 - 24 months and become a very small minority.
Evans estimates the following smartphone app usage for Facebook (based on Q3 data above):
Below is a chart from Evans showing the relative growth of Facebook access on the various mobile platforms from September 2011 to September 2012:
Assuming these numbers are accurate you can see the reversal of positions of the iPhone and Android since last year, which makes sense. However the larger point is that a majority of Facebook users now access the site via mobile.
Facebook has argued that mobile is ultimately a much larger revenue opportunity than the PC. The following verbatim Facebook remarks come from the Q3 earnings call transcript:
Social and paid-search ad platform Kenshoo came out with data today that argue the percentage of ad revenue coming from mobile is now up to 20.3%.
Facebook is expected to generate roughly $1.5 billion in overall revenue in Q4. Not all of it is ad revenue, however. Roughly $260 or so million would be attributable to mobile if the 20.3% figure holds and the forecast is correct.
Google sees lower CPC prices on mobile paid search ads but better performance on mobile devices vs. the PC. However Facebook is experiencing the opposite phenomenon, according to Kenshoo. It sees higher mobile prices but lower engagement vs. the desktop.
I had an interesting experience this past week with my 13-year old daughter, which illustrates the challenges but also the opportunity for Windows Phones. One of several smartphones I have is a blue/purple HTC 8X (Windows Phone). The phone offers Beats Audio integration. It's a very attractive handset and looks very much like a Nokia Lumia device, only not as heavy. (Nokia is not too happy about the close similarity.)
My daughter is currently a feature phone user and really wanted an iPhone 5, which my wife and I cruelly denied her. But she spied the 8X on my desk and really liked how it looked. She also had seen (on Hulu) the relentless Microsoft Windows Phone ads -- "as unique as you are" -- and was parroting some of the ad copy/dialogue to me almost verbatim. (The campaign must be working.)
She asked for the phone and I agreed that she could have it. I gave it to her to try at home on WiFi to make sure that she was really interested before I went through the trouble of changing carriers and so on. I suspected the OS might throw her; she's had an iPod Touch for several years and we've also had several Android phones. She's familiar with both operating systems but hasn't ever used Windows.
Another factor: she was intrigued by the Microsoft Surface tablet, which she saw at a friend's house. The colorful and different look of the UI appealed to her. The Windows "metro" design -- I can still use that term even if Microsoft cannot -- on both devices got her attention.
Then she started playing with the phone and found out that some of the key apps that she and her friends routinely use weren't there. She wasn't thrown by the metro UI and "live tiles," as I expected. Rather it was the fact that Instagram, Oovoo, Snapchat and other apps she uses were missing. She was particularly annoyed by the pseudo Instagram apps that appeared (e.g., Instagram blog). After discovering that these beloved apps were missing she rejected the phone.
Had those and a few other critical (in her mind) apps been present, she would have kept and used the phone. I told her that Instagram would eventually come to Windows Phones as would other missing apps. That didn't satisfy her.
Windows Phone now has 150,000 apps; however as indicated by the above it's still missing many key apps. For example, as part of its antitrust argument against Google Redmond says it's being unfairly denied access to YouTube metadata. It's makeshift YouTube app is deficient in fundamental respects:
Google has refused to allow Microsoft’s new Windows Phones to access this YouTube metadata in the same way that Android phones and iPhones do. As a result, Microsoft’s YouTube “app” on Windows Phones is basically just a browser displaying YouTube’s mobile Web site, without the rich functionality offered on competing phones. Microsoft is ready to release a high quality YouTube app for Windows Phone. We just need permission to access YouTube in the way that other phones already do, permission Google has refused to provide.
If Microsoft can manage to get the key/top apps onto Windows Phone it will be able to attract buyers who like the different look of the UI and/or who may be attracted by the aggressive subsidies offered by carriers. The 8X is available for $49 with a two-year contract at Verizon, for example.
However until these apps (enough apps, key apps) are there would-be users, such as my daughter, won't bite. Microsoft can break the "catch-22" of Windows Phone app development through developer payments and incentives, which it's trying to do. The company needs to identify all the "necessary apps" and make sure those are built for Windows Phones.
Beyond this the current messaging isn't really successful in attracting buyers, notwithstanding my daughter's ability to parrot the commercials. Microsoft needs build messaging around three ideas:
Finally, ideally, there should be something about the hardware and software (beyond the UI) that is different. Microsoft argues there are already such things (e.g. Kids Corner). And the colorful phone cases offer an approach taken by Nokia and HTC. A better camera (i.e., Lumia) is another.
Yet these things aren't quite enough. There needs to be a highly visible feature or dimension of Windows Phones that truly isn't present on iPhone or Android handsets. Right now, nothwithstading the nicely designed metro UI, Windows Phones continue to seem like "wannabe" devices that are still playing catch up to other smartphones.