Context is King: Google Parameterless Searches

In the very near future, you may be able to perform searches at Google without bothering to type or speak a query. Instead, you might be able to just shake your phone, or hold down a button for a certain amount of time, and tell your phone something like “search now”. Known as parameterless searches, this type of search can depend upon the context within which the search is performed.

Screenshot from the patent showing someone shaking a phone, context information for that search, and a search result

For instance, imagine being driven to work at 50mph, and you shake your phone. It tells you that there’s congestion ahead, and offers an alternative route.Or it shows you a map with color-coded traffic information for different streets nearby according to traffic conditions. Or, you may have an appointment with a client made by email and included on your calendar, and you want to find and check the email to make sure that you have the right phone number. It could show the number and offer to make the call on your behalf. If you regularly take a train at around 8:00 am on weekday mornings, shaking your phone at 7:50 am might trigger a realtime schedule for the rails.

Context information for a parameterless search could include things such as:

  • Time/date information,
  • Geographic location information,
  • Calendar information,
  • Rate of speed information, and
  • Device activity information (like the sending of emails for different purposes).

The patent is:

Providing results to parameterless search queries
Invented by Sumit Agarwal, Vic Gundotra, Alex Nicolaou
Assigned to Google
US Patent 8,478,519
Granted July 2, 2013
Filed: August 30, 2010

Abstract

In one implementation, a computer-implemented method includes receiving a parameterless search request, which was provided to a mobile computing device, for information that is relevant to a user of the mobile computing device. The method also includes, in response to the received parameterless search request, identifying with a digital computer system one or more results that are determined to be relevant to the user of the mobile computing device based upon a current context of the mobile computing device. The method further includes providing the results for display to a user of the mobile computing device.

Context information could more specifically include things such as:

  • Geographic location,
  • Weather conditions,
  • Nearby businesses,
  • Volume of ambient noise,
  • Level of ambient light,
  • An image captured by the mobile device’s camera,
  • Rate of traveling speed,
  • Time and date information,
  • Calendar appointments,
  • Recent user activity,
  • Habitual user activity

The current context can be determined by data and sensors that are local, or even remote to the mobile computing device. For example, traffic speed and congestion might be determined by sensors in the phones on the road ahead.

If someone has been checking a developing news story via their browser or a news application, and checked a few times, a parameterless search may indicate a desire to see an update to the story.

If someone is driving down the road, and following a news story, a parameterless query could trigger responses to both.

This system is self training, and can learn about different categories of information that you might be interested in, based on contexts. For example, if the results of a parameterless query ends up showing a list of local restaurants, and you end up choosing one of them and driving to it, this system learns that in contexts like that, a set of restaurant results is an appropriate response (and it may learn about your taste in food as well). It can take learn about other user actions as well:

For example, if the user is provided with a recent update to a blog that the user frequently reads and the user sends a link to the updated blog posting to his/her friends, the user sending the link can be recorded as user behavior data in the user behavior data repository of the mobile computing device and used to provide results to future parameterless search requests.

The patent tells us that this parameterless search query approach isn’t limited to mobile phones, but can work with tablets, laptops, car navigation systems, personal digital assistants, or desktop computers.

I’m wondering how Google Analytics might indicate that a search leading to a site might indicate referral information – no query?

Share

35 thoughts on “Context is King: Google Parameterless Searches”

  1. Thanks Bill – I found this to be a particularly interesting and telling update. In relation to the last question, probably not provided. How long until the concept of (and data on) keyword traffic is basically dead? Will that be a positive thing? less data but perhaps these types of changes will require marketers to focus more squarely on establishing their biz for relevant themes and niches

  2. It sounds exciting and a challenging task for the search engine. But I believe this would cause some privacy issue as this gives Google a lot more information about the user. Also there is the possibility of getting irrelevant information at times when you need something specific.

    Google Analytics showing (not provided) is already a big loss of keyword rich data to webmaster. Guess what impact will this have.

  3. The word I would use is non-parametric like medians, ranks, and Kolmogorov–Smirnov normality tests since Google is supposed to be so engineer geeky. As opposed to MS Nerdy.
    The kind of searches this guy is writing about are not non-parametric (or “parameterless”) either but context sensitive searches which are nothing new. So many of these articles are written by people who obviously have no background in math, stats, or programming.
    I’m not sure what would be a good non-parametric example in the context of search analogous to distribution free as a property of data. Maybe query free, predictive, static, ontological, …

  4. The context information is interesting. There’s household equipment able to communicate to each other already, so I’m wondering if Google will somehow incorporate that. Running late? Shake your phone for “quick wake up settings”. Shower turns on, coffee brews, and 15 minutes later your car starts warming up.

  5. Hi jgury

    I didn’t make Google’s patent up out of my imagination. Read it. Then if you would like, argue with Google’s search engineers about the terms they decided to use. Not me.

  6. Great direction, great analysis. This is the next development step of Google Now so it seems, and not actually very far out.

    Even if we leave out the further spin (like Jared suggests) of home automation and focus just on searches, it has a lot of positive and negative implications – negative for sure everything related to reliance during times of outages, huge privacy issue potential and also security.

    Simple example: I search from home vs. from work, and I am searching for completely different information. It would be great to have google switch the personalization of my results based on time plus location. search for SED at home might be stocks or some organization, sed at work or home at work hours might be the linux tool. That would be great as I have been stuck a few times with not helpful searches in personalization.

    This is one step more in the direction of personalization of results – clearly not making it easier for SEO, but it confirms the old message – focus on the customer, not on SEO.

  7. In regards to referral information being driven to a particular site, it seems as if most of what these parameter-less searches are doing, is pulling a specific data set. Like an API. A site wouldn’t get pulled in itself, but might show it’s name for reference for the data pulled. Weather info from weather.com, traffic data from traffic.com (making this stuff up) but I think you get the point.

  8. This is very interesting news indeed. I think there will always be a spoken or typed keyword, but for local business – hmmm. Very interesting. Thanks so much for the info, Bill.

    Mel

  9. Are there any instances currently where parameterless searches are conducted? Siri is so bad that I know she doesnt do it but is it possible this is already in play on a smaller scale?

  10. Fascinating stuff Bill. It seems like wearables and Google Glass in particular, would be a huge benefactor as Google reduces this patent to practice. The trade-off is privacy vs convenience and whether the additional convenience justifies the sharing on way to much information. Like the current events in the news (US primarily) where the trade-off is privacy vs security, it seems like the methods discussed might meet some strong resistance given how many people are reluctant to trade privacy for security much less convenience. Europe will be even a bigger challenge for acceptance/tolerance fo these kinds of methods. Nonetheless, this kind of search is coming for those that willing to share every detail of their life. I guess Facebook, Twitter and Four Square are proof points that this kind of transparency scenario is plausible for some. Thanks for sharing.

  11. Hi David,

    Google has referred to and hinted at parameterless search in the past, in places like the predictive algorithm behind Google Now, for instance.

    I do believe that things like the Google Knowledge Base pose just as much, if not more, of a risk to searches defined by keywords. Google’s been moving away from strict keyword matches for a while now, with synonyms and concepts playing a stronger and stronger role in results that are returned when someone write keywords into queries. The existence of “related” terms (as defined by co-occurrence in places like Google’s Phrase-based indexing, or word relationship patents) on pages in response to a particular query can cause those pages to be booted in search results. Anchor text that use “related” terms in links to those pages can also cause different amounts of PageRank to be passed along.

    The days of just matching keywords to pages is over. As you note, marketers should be focusing “more squarely on establishing their biz for relevant themes and niches” (or at least relevant and related topics).

  12. Hi Ehsan,

    Google has never hidden the fact that they will use information about the context of our searches to work to improve those searches. Of course there’s the possibility of getting irrelevant information when you need something specific. The response to that is to actually perform a search using queries that are a better fit.

    The “not provided” results in analytics are supposed to be to protect the privacy of people searching. If Google is going to protect the privacy of searchers, there are going to need to be “not provided” type results.

  13. Hi Jared,

    It is amazing the kinds of capabilities that mobile devices might bring us in the future. The parameter search patent hints at giving us the ability to actually control things like this, like the example of noticing the scheduling of a phone appointment, and offering to dial the number, based upon that context. I suspect at some point though, that our “home” computer will manage things like the making of coffee, preparing breakfast, and turning on the shower. :) It might do that by communicating with the computer in our mobile device.

  14. Hi Andreas,

    Thanks. I agree that this description of parameterless searches does lead down the path towards intelligent assistants like that described in predictive algorithms that might help define Google Now.

    The things you mention about personalization are things we should be thinking of now anyway when doing SEO – recognizing that the searchers we create pages for may have very different reasons for searching depending upon whether they are at work, or at home. The types of devices that they perform searches on might differ based upon those contexts as well – even when searching for the same things. I might, for instance, do some exploratory searches on my phone at the store, and then switch over to my desktop computer to order something that I did see in the store.

  15. Hi Zander,

    I wondered that myself when starting to read the patent, but then came across the example of it noticing that you may have been following along a specific storyline, and as a result, Google might want to provide updates to that story. I could see that resulting in the display of a set of related search results.

  16. Hi Mel,

    I could see Google showing off local results for specific types of businesses based upon user behavior and user history. Someone is driving most of the morning, and it’s getting close to lunch time. Our smartphone user often stops at any thai restaurants he finds for lunch. A parameterless search might show off a set of local results for nearby thai restaurants. :)

  17. Hi Bonnie,

    I don’t know of any at present, though it’s possible that some have tried in the past. The original Siri, as described in a patent from SRI may have included that capability, but when Siri was released by Apple, it didn’t include a lot of the capabilities that were described in that original patent.

  18. Hi Rick,

    Google did release Google Now on a mobile device, and the kind of parameterless searches described in this patent would work there as well. This is the kind of feature that would probably work well with wearables like Google Glass, the smartwatch that Google has released a couple of patent filings on, and others. Google has published a couple of patents about how user behavior and location based information associated with users might be quarantined, with some aspects (such as personally identified information) expunged from that information. See, for instance, http://www.seobythesea.com/2013/02/google-patents-identifying-user-location-spam/

    It’s going to be interesting seeing how our ideas related to privacy and convenience evolve.

    For instance, there’s a lot of information that is publicly accessible, but has been so difficult to access that it has been treated as if it was protected. For example, law suit judgments as usually kept as publicly accessible records at local county courthouses. They often have social security numbers printed on them. How does that change when those records become available online?

  19. good article – I was just wondering where you got the illustration that you used above? Did you make it yourself? Because I can’t see a source and I want to use it in a presentation about the exact topic. Would be nice if you could either tell me where you got it from or allow me to use it (of course, I would state you as the source then).

    Lovely regards,

    Toby

  20. Hi Tobias,

    The image is a screenshot from the patent, itself. As a publicly filed document, it’s not under copyright. So, you can use it in your presentation regarding this patent on parameterless search, and should probably credit either to the patent itself, or if you would like, to the patent via me (if you want to give me credit from digging the patent and image out of the USPTO database. :)

  21. Hey Bill just popped back in to say hello.

    Thank you for being nice about links on your site.

    Contextual search would be awesome. Larry used to joke we’d have a microchip implanted in our brains and whenever we had a question / search, it would automatically pull up the relevant information.

    -V

  22. In short, any website that exists on advertising revenue and offers the above mentioned data is about to lose a large chunk of their revenue.

  23. Shake my phone a certain way? That sounds like way too much work… Oh wait, I don’t even have a smart phone! lol

    Thanks for the update, Bill.

  24. Great post. Everybody believes to make his own personnal searches. And Google insight reveals that everybody researches the same thing influenced by same context : news, product launch, weather forecast, season searches, even flu-related searches as Google Flu Trends reveals it.
    Forums are a great tool to reveal these commun searches. Sometimes you do a search and found that this question is already archived in a forum thread. Because this “newbie question” is a very commun search indeed that frequently come back in a particular context.

  25. Things like what this blog post covers are what dreams are made of. I guess there’s not much reason for being pulled over for using your phone by shaking a phone instead of having to mess about with it just to get hands free properly working. It wouldn’t surprise me if this was out sooner than later with the way Google is going!

  26. Hello Bill,
    Completely new information. Through all such patents and experiments, one thing is clear; Google is trying to read users’ behavioral pattern or requirements on different contexts etc. The prime focus is to assess the thought pattern. But basically as human beings, our thought pattern is full of randomness. So, judging one’s mind in a particular order of behavior or habit is really erroneous. Our living style is highly unpredictable. So, it is very difficult (of course not impossible) to put in a practical application. Till then I think Larry’s idea of implanting a chip will be the best shot.

  27. I love learning about the future of search–I’ve written about it once or twice myself–but the idea of reading an entire patent does NOT sound exciting! Thanks for going through it and giving us a summary. If I ever talk about this crazy new “parameterless” search idea, I’ll remember I found it here.

    To your last point–it almost sounds like something the new “Universal” Google Analytics would handle. Those tend to handle search-engine-agnostic events (like a check-in at a gym) and a parameterless search might fall under that category. Because there would be a couple conditions that affect what result you get, the “single keyword > 10 blue links SERP > google/organic” attribution wouldn’t really be appropriate. Will be interesting to see what happens!

  28. I cannot believe patents like this are being issued. Smart phone have been doing this to some extent years before this patent was filed. This is not new and certainly should not be patentable. This is just another instance of a patent where it will a) limit functionality and innovation in the marketplace, and b) create years of litigation.

  29. It actually almost boggles the mind how they would achieve this in a way that worked. One thing I did think of that would be cool was if a mobile was picked up the in dark and given a shake, the flashlight (from camera flash) would come on. Yet this is a phone function more than a search function.

    While I’m sure that Google will have answers to what can be achived, like with Google Glass, I’m personally a bit concerned about privacy issues. It’s actually an indictment on society as we are these days that what initially causes privacy concerns tends to fade away and become accepted within a short period of time.

  30. I can’t see how this will be more beneficial than say being able to voice type a search query into a mobile device. When I search I seek specific information, and I do not see this being something I will received. Plus I don’t feel comfortable with Google knowing I’m traveling down a specific highway, at a specific speed, at a specific time of day. Way to invasive.

  31. I think that this is just a step in the direction of always on. Many of the sensors we’ll wear and carry will always be on, Google (or any other service) would try to understand our needs even before we do.
    Guessing the easiest would be to learn our habits and slowly work our form there.

    Privacy? Different subject…

  32. The very term “Search” is wrong. It’s not about search. It’s about life assistance. My parents might think about search as “We solved the argument about who that character actor is from that 70’s flick” or simply what movies are playing and when.

    I think of “search” as: When did I last speak to him? What were her kids’ names again? What is my average time leaving the house in the last year? Does my wife think this is the right food to be buying for my nine month old? And yes, most definitely, is there traffic ahead?

    And here’s the catch: I don’t want to type. I don’t want to speak. I don’t want to shake. I want the “machine” to simply know. Whether that machine is a shakeable cell phone, a car gps, or something else, I don’t care. I just want to know. And I don’t want to ask.

    It’s probably something like google glass. “It” probably needs to see what I see, hear what I hear, and know what I already know.

    “It” is simply me. The machine is an extension of me. Life is too complicated. I need an external hard drive and a port to all the knowledge in the world. Google is just trying to make the best software to go from in here to out there. I welcome it.

    Privacy is irrelevant. The human compact will keep it in tact to some degree for an extended period of time. Think about when those idiots at facebook started publishing what I was buying without my permission. They got sued. No one was ready for that. It stopped. It’s tacky. They still know what you’re buying. Now they only use it for themselves privately.

    And Len, Google already knows where you are and how fast your going. They and the 50 other companies in your pocket know way more invasive things than that. But they only care so they know what you want to buy and when. Embrace it. It’s either that or some idiot calling you in the middle of dinner to sell you timeshares from some place you went ten years ago and didn’t even like.

    What Google is giving me is worth sooo much more than what they want. They’ve given me access to everything. All they want is for me, when I want to buy something, to ask them who to go to. No problem.

    Thanks Bill for the info. I’m happy to know we’re one step further to my plugged in life.

    disclaimer: More than 1/2 of my IRA is in GOOG stock. Why wouldn’t it be?

  33. Yes, you are right that content is king. You can absolutely increase your visibility and boost your rankings if you will create high quality content.

Comments are closed.