Google Experimenting with Specific Answers Without Any Search Results

[engine]

Google is showing an answer on a search page for certain queries, but there are no other search results being shown. ie, ten blue links gone. I know, ten blue links are long gone, as such, but you know what i mean.

For example [google.com...]

[searchengineland.com...]

It appears to be for a limited number of search queries, such as the time, calculator, unit converter.

Google's Danny Sullivan says,

For calculator, unit converter & local time, we’re experimenting with a condensed view to further speed up load time. People who search for these tools rarely use full search results, but the results will remain available for those who want them via the "Show all results" button.

[twitter.com...]

[Shepherd]

So, google has a search engine (google is not a search engine). A what point does a service google provides become not a product of their search engine? When search results are no longer shown as a part of the service provided?

This may seem like a pointless discussion now, but at some point if/when the monopoly mob comes running with their pitchforks little things like this will matter.

[glitterball]

Google obviously feel that, legally, they would not get away with just lifting sports scores from elsewhere.


Obviously it's difficult to have a logical discussion with so much personal assumption being thrown around as fact.

Fair enough - but it does beg the question - what other snippets/news/data/facts do Google pay to license and reproduce?

[brotherhood of LAN]

Fair enough - but it does beg the question - what other snippets/news/data/facts do Google pay to license and reproduce?

Isn't it relatively obvious from the structure of what's served?

The cruft from web pages tends to be blobs of unstructured text and lists, which is seemingly summarised from a consensus of other web pages. They do feeds like sports results. They do specific answers like in this thread.

Maybe there's some thinking cross over from their knowledge graph stuff (and the lifting of text for summaries), which appears to be entirely separate.

[glitterball]

Isn't it relatively obvious from the structure of what's served?

Not to me - I looked up some UK-oriented sports queries, and I can't see any credit to any sources, so how would I know if it's paid for or not?
Forgive me if I'm missing something, but I would genuinely like to know how you can tell if Google have paid for the data or not.

[TravisDGarrett]

As for time, there is no need of feeds of any kind. The clock of the server, and knowing the time zone of each cities is enough. Same for sunrise / sunset which can be mathematically computed.

As for sports results, Google can also simply put a guy in front of TV screens watching sports, and just reporting the score ...

[Shepherd]

but I would genuinely like to know how you can tell if Google have paid for the data or not.

Of course, there's a lot of trade secrets out there I would like to know also...
That said, if you are seeing data displayed without credit to a source there's a high probability that it was licensed from a company that provides that data or google collected it themselves.

[glitterball]

or google collected it themselves.

You can see my confusion - do you mean Google wrote it themselves?

As for sports results, Google can also simply put a guy in front of TV screens watching sports, and just reporting the score ...

That was my presumption too, but as per the post earlier in this thread, it seems that is not the case:
[webmasterworld.com...]

[TravisDGarrett]

Google wrote it themselves?

Google's AI (the two magical letters since some months), can certainly browse the internet, read things her and there, and write its own content.

[brotherhood of LAN]

I would genuinely like to know how you can tell if Google have paid for the data or not.

Maybe they have mentioned it in the past, I don't know. Google do have a link to many of the examples here though, outside of knowledge graph, product feeds, news etc that we're used to.

[support.google.com...]

Stock market, sports and measurements all get singled out as part of their product.

[glitterball]

@brotherhood of LAN : Very interesting how they have divided up those categories in your link. Seems that the first one could all have an open source origin, whereas the other two may require some kind of licensing.

[glitterball]

Google's AI (the two magical letters since some months), can certainly browse the internet, read things her and there, and write its own content.

Blurred lines indeed.

[Shepherd]

do you mean Google wrote it themselves?

"wrote" would infer creation, data (specifically sports scores) is created by an event, collected and then disseminated. Another example: street view, google did not create what is in the images they only collect and disseminate.

[brotherhood of LAN]

Seems that the first one could all have an open source origin

Yes, that's how I understand the last two are provided. For the former [support.google.com...] ... 'Facts and info' is their knowledge graph.

For 'specific answers' wrt this thread, ' time, calculator, unit converter ' as mentioned in engine's original post here are not showing any organic, and perhaps if Google deems them beneficial, the others in that list will follow, like when public holidays are etc, which Google already shows additional data for.

[glitterball]

Another example: street view, google did not create what is in the images they only collect and disseminate.

Not really, Google did create those images, so they are Google's original work. There are exceptions, e.g. the copyright of the lighting of the Eiffel Tower, but generally photographs are considered original works.

[Shepherd]

Read all the words, google did not create what is in the images.

[glitterball]

Read all the words, google did not create what is in the images.

Yes, but when it comes to the actual digital media that is being shared, in that case Google did create the original work.
Therefore, Google would own the rights to those images.

So, in your own words, they did not just "only collect and disseminate".

[Shepherd]

Yes, in a completely unrelated topic, it is possible to create an original work during the collection and dissemination of data.

[EditorialGuy]

Google's AI (the two magical letters since some months), can certainly browse the internet, read things her and there, and write its own content.

You and I can, too. If I were to write a page about the New York subways, where do you think I'd get information on subway lines, fares, etc.? My eyes would be "scraping" that information from the official transit site. It's called "research."

[glitterball]

My eyes would be "scraping" that information from the official transit site. It's called "research."

This is the crux of the issue: where do we draw, or redraw the line?

To use your analogy, if your ears "scrape" a number of notes from a melody, can you reproduce them and call it research?
At what point is it plagiarism?

I think that a lot of the members here are concerned that this is just the tip of the iceberg.

Imagine this possible future scenario:
A user asks Google "What is the Capital of France?"
Google then returns "Paris" as the answer, but instead of showing any website links, it returns a list of 'related questions':
What is the population of Paris?
What attractions should I see in Paris?
Where should I stay in Paris?
etc.etc.
All of which would return a further short answer along with lists of hyperlinked related questions and answers.

Now, you may very well argue that this is of benefit to the user, however at what point are we destroying the ecosystem to give the user what they want in the short term?

Another question that I would ask, related to so-called AI recompiling information from the web into new sentences and paragraphs is this:
If I was to write a simple program that took whole texts from elsewhere and replaced random words using synonyms and re-ordered a few paragraphs, at what point would it stop being plagiarism?

[EditorialGuy]

All of which would return a further short answer along with lists of hyperlinked related questions and answers.

I've seen Web sites and apps that were built on that model. They haven't done very well, probably because the interface is so artificial.

Ideas and facts are a dime a dozen. Presentation is what makes the difference between a bestseller and a flop.

[londrum]

google are always going to be constricted by the size of the snippet they can reproduce. presumably that's why they upped the character count for meta descriptions a while ago. eventually they are going to have to create more of their own material. that's the next thing that's coming. (imagine if they were allowed to buy wikipedia - all of our traffic would disappear overnight!)

[EditorialGuy]

google are always going to be constricted by the size of the snippet they can reproduce. presumably that's why they upped the character count for meta descriptions a while ago. eventually they are going to have to create more of their own material. that's the next thing that's coming.

Remember the Frommer's purchase? Or the Zagat acquisition? They've since dumped both.

Google does much better as a search engine, a hosting service for third-party content (YouTube), and an advertising firm than it ever could do as a media producer.

[EditorialGuy]

Addendum: Let's not forget, too, that Google profits from third-party content via DoubleClick, AdSense, etc. Google has a vested interest in a healthy Web-publishing industry.

[brotherhood of LAN]

Addendum: Let's not forget, too, that Google profits from third-party content via DoubleClick, AdSense, etc. Google has a vested interest in a healthy Web-publishing industry.

but instead of showing any website links, it returns a list of 'related questions':

I feel these are wide of the mark. If I stop someone on the street and ask them the time, and they start talking about the time difference in another continent, you can get the feeling they're not satisfying my query.

Google already shows 'related' searches and has done so for a very long time.

The answers provided in the context of the original post are for matter of fact answers, unequivocal. It's fine to take exception with Google being the deliverer of that information, but if you create a reasonably decent query parser then you/anyone would be able to deliver the same information, on a website/search engine/app or whatever.

[engine]

Google's Danny Sullivan now says the experiment is over, and I note, he adds, "for now."

I am sure it'll be back, and DS says, "The team will look at improving when and how it appears."
[twitter.com...]

[MrSavage]

LOL. Google: "We bring the web to you so you don't have to go there".

I love to dissect PR press releases. So Danny says working on improving "how it appears". Right. As in what? No links is no links. A calculator is a calculator. It's pretty much a blank page as it was, so what is there to improve on in terms of how it appears? Google struggling with how one item (two if you include the link to show paths to the scary world wide web) appears on a page? Awfully puzzling and I tend to just use it an an illustration of slinging BS.

In terms of "look at improving"? I mean again, it's one page element! Improve it how? Make it more robust, thus completely removing the idea that anyone would need to click that scary "see more" button? Blinking lights? I mean improving the "what time is it?" result page? Improving a calculator? Makes no sense.

They did it and can't take it back. For the first time they all but blocked the internet behind a button. Shameful quite frankly. Make a new product that hides the web. When Google "search" hides the web behind a button, most people who launch things to the world wide web should be troubled by that.

[thedonald123]

When Google "search" hides the web behind a button, most people who launch things to the world wide web should be troubled by that.

The way this became public was strange from the beginning. Google is always running experiments in the SERPS to a small segment of users. You can see them reported on Twitter by the SEO community and Barry reports them on SEL. There is constant reporting on experiments people notice with screenshots etc. Google never announces an official change until the experiment is over and they roll it out publicly to all users.

This whole public release of a "experiment" feature and public rollback is highly suspicious. If it was an experiment why test it on everyone?

Seems to me someone somewhere higher up realized the huge legal ramifications of changing the official Google.com from search engine always displaying links to other websites to a stand alone website. Once Google.com is outed as a website, which they are and technically always have been, then the media and the public might start asking questions which Google would prefer not to answer just yet, e.g. where are they stealing their content from. Is it licensed?

[EditorialGuy]

where are they stealing their content from. Is it licensed?

We touched on this earlier in the thread. Some feeds are licensed, some information may be gathered via APIs that any Web site is welcome to use, and some direct answers are almost certainly generated internally (e.g., the answer to an arithmetic problem or an imperial-metric conversion). The notion that Google and other search engines are "stealing" the current time or the answer to "2 + 2 =" is pretty farfetched.

[thedonald123]

@EditorialGuy yes, I wasn't being clear. I meant the new features when Google uses snippets from my copyrighted content to fully answer related questions in the SERPS. Or on voice search where there is attribution of the source website. What's with the attribution? It's as if Google is saying the copyright owner by not blocking Google in robots.txt has agreed to license their content with a Creative Commons license and all that is required is attribution.
Why is no declaration of license needed? When does it stop being "fair use"?
>>where are they stealing their content from. Is it licensed?

[EditorialGuy]

I meant the new features when Google uses snippets from my copyrighted content to fully answer related questions in the SERPS.

Every example I've seen includes a link to the source, so that's your attribution and "fair use." And for some of us, those answer boxes and featured snippets are bringing in considerable traffic.

Still, site owners who don't want their content appearing in featured snippets can simply use <META NAME="ROBOTS" CONTENT="NOSNIPPET"> or <META NAME="GOOGLEBOT" CONTENT="NOSNIPPET">.

[support.google.com...]