Two search technologists previously from Yahoo and Powerset have teamed up to create an app designed to turn searching on its head. Jybe, founded by Arnab Bhattacharjee (who previously ran the Yahoo! Search Technology team), Tim Converse (ex-Powerset) and Tuoc Vinh Luong (the VC savvy), spoke to SEW over the phone about their new iPhone app that aims to surface suggestions of what movies to watch, restaurants and dishes you'll like and books you will find interesting.
The Problem with Explicit Search on Mobile
The big difference is that this is search technology reversed - from explicit, query led search engines to implicit, suggested discovery orientated "search agents". Another phrase for this conceptual approach could be "digital concierge". Siri and Wolfram Alpha are the closest things we currently have to the idealized concept of search agents currently, however 'apps' are truly the forerunners of this concept - when Steve Jobs famously said, "When you look at ads on a phone, it's not like a desktop… On a desktop, search is where it's at. But on mobile devices, that hasn't happened. Search is not happening on phones; people are using apps. And this is where the opportunity is to deliver advertising."
Whilst the numbers now around mobile search and mobile search ads do not necessarily vindicate Job's "shot across Google's bow", but nonetheless a certain prescience remains. Searching on mobile devices via explicit query-led search is a fairly awkward experience. New signals for search engines such as location data and even the quality of data connections is not only improving the mobile search experience but changing it too. Explicit search queries tend to have a fewer number of keywords on mobile devices than on desktop devices. This may be because of users' dislike of typing on a small screen, being on the move and so having less time, or because users implicitly expect the mobiles to be able to calculate their intention. Equally, mobile search could be so successful now that users require less queries or vertical search is so well served by mobile apps for niche needs that less explicit search is required on mobile. Either way, the data challenge mobile search engines ultimately face is that user led direct intention signals (i.e. the query string) are diminishing whilst the devices themselves hold the technological capability to send many many more contextual signals.
Simultaneously, monetizing mobile search seems to threaten the quality of the explicit search experience itself because the lack of screen estate means that organic results are pushed further down the page in favor of ads.
Improving Search on Mobile Devices
With this in mind, Jybe has jumped into an interesting opportunity to tackle the potential around mobile search. If you don't already know what books to read, films to watch or what to eat on the menu, explicit query led research into this is likely to be extremely tedious and frustrating on a mobile device.
Personally, whilst I think Mobile Search is generally good on Google in particular, I can find it frustrating when mobile device optimized local results are not returned for queries that, to me, have the same intention - the query 'movies' or 'films' can often bring up different result sets, when really, from a mobile device, I just want to know what films are on now and where a nearby cinema is. Equally there can be an annoying specificity required needed to differentiate between say, actual web destinations for hotels versus aggregator and review sites - which is a problem I find especially apparent in mobile paid search ads. Often in these instances I prefer to search via apps - Google Maps for instance seems infinitely easier to use to browse quickly through local options than "the 10 blue links" display, although this can suffer from keyword targeting issues sometimes. However, in general, when mobile search results, organic or paid, really work for me is when I can take an action directly from the result, whether that is a clickable phone number to immediately place a call or the second click is a list of film times at the closest cinema.
To get Jybe up and running you simply have to input your preferences via a quick questionnaire, much like how it works on Netflix or in the Alfred App (recently acquired by Google) and you can plug in other data sources too such as Twitter and Facebook data, which Jybe data mines to calculate how much your tastes differ or concur with your friends and also across the entire social graph to find common personas who share similar interests as you. However, where Jybe really excels is the actions you can then take based on the recommendations it provides. With Jybe you can do all of the below, all within the app:
- Restaurant - Call the restaurant, Make a reservation through Open Table, Find a recipe and get directions.
- Book - Download to iBooks, Kindle or buy physical book from Amazon.
- Movies - See cinema showtimes near you, buy tickets via Fandango, get directions to theater OR watch the trailer, add the movie to your Netflix queue or buy it on DVD from Amazon.
You can also save suggestions to a to-do list and, in the case of restaurants, the menus are reformatted such that it can instantly suggest a dish that you will probably like at the restaurant you are at. By contrast, its closes competitor was Alfred by Clever Sense (now Google) which offers very little in the way of actions.
Search Agents Can Expedite Customer Conversions
This is where Jybe strikes me as having set out on the right path. The battleground for mobile search mindshare is not necessarily in simply presenting the right results more often than not, but in being able to expedite the next action required to fulfill on the user intention. The faster a user can make a call, buy a ticket, make a reservation or confirm where they are headed, the more useful the entire mobile search experience is, regardless of whether that search was performed explicity or implicitly. Where relevancy is paramount on desktop search, serendipity in mobile search elicits forgiveness for any previously irrelevant experiences. The difference in experience here is crucial to understand - whereas for the sense of relevancy to manifest is a relatively cerebral process in the user that simply leads to the next click, for serendipity to manifest, it has to be tied to an action - something has "to happen" for it to be called a moment of serendipity. And the point is, we may not all fully realize it yet, but what we are really demanding from the mobile search experience is to take more action, because more action equates to more fulfillment, both in the cold transactional business sense of fulfillment (e.g. completion of customer order forms) and in the, warm, fuzzy, emotional sense (e.g. I wanted this, found that, and these things happened).
Whilst Jybe currently can only expedite conversions for services that offer an open API and is heavily reliant on those services remaining open, it's easy to see that there could be many more conversion, or "consumption points" as Bhattacharjee put it, in the future. Partnerships or the availability of more open APIs could see us auto registering for deals and offers, app or music downloads, adding to YouTube playlists, bookmarking on Hulu or even ordering specialty foods and drink from Whole Foods. In the far future, digital concierges may be stocking our fridge, ordering taxis and recommending diets if the digital permission system of semantic markup and open APIs allow it.
Data Sharing Permissions Herald Era of Connected Marketing
The crux of my argument for online marketers is to urge you to explore the opportunities among emerging Search Agents and Digital Concierge technologies, is that success criteria for mobile search should not be focussed on the beginning of the user journey, but the end of it. Mobile search technologies need to fulfill the customer intent to transact and consume as quickly as possible - and they can only do this in partnership or with support for open protocols.
Ultimately it is all about "data permissions" (orchestrated via apps and APIs and validated by the user), which is the fundamental premise behind Connected Marketing - a concept which is still very much ahead of its time. Services like Jybe should remind the most committed online marketers that the future of our business may not lie in our actual products or services but in how we structure all the data around them in such a way that other services can use them in return for delivering conversions even quicker and more cheaply to us. If you can get your head around what a "connected business" really is, you'll have a eureka moment about what "connected marketing" could really do for lead generation.