Forum Moderators: phranque

Message Too Old, No Replies

Search Engine Philosophy Regarding Answers to Questions

         

brotherhood of LAN

6:45 pm on Oct 2, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



A question I'm having to grapple with as per title.

Many search engine queries can be evaluated and reduced into a simple 'universal truth' answers, or at least an authoritative one.

Examples include weather, currency conversion, famous people/places/events. Often they appear on Google and Bing as knowledge panels or straight up answers in large text at the head of search results. e.g. "David Beckham age", "GBP to USD", "weather in X" - I'm sure you've seen plenty examples in their various incarnations.

I'm interested in what your thoughts are in regards to this intervention of a search engine in providing an answer. Is it common sense to satisfy the query, or is it the job to serve up web pages that satisfy those kinds of questions? Not interested in SE bashing, more interested in an academic opinion of it as to what are the bounds in what a search engine should operate.

Also interested in if you see this changing in the near future given voice search and anything else emerging.

iamlost

11:39 pm on Oct 2, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month




what your thoughts are in regards to this intervention of a search engine in providing an answer. Is it common sense to satisfy the query, or is it the job to serve up web pages that satisfy those kinds of questions?

It really is a matter of definition applied to capability :)

At it's most basic a search engine is a software information retrieval system.
The term, as commonly used, means a software system that takes a (usually) text input aka query and searches either or both a saved index or the Internet/WWW in real time for valid answers/responses that it subsequently displays usually in some method of relevance.

The first popular SE was Archie (1990), which searched/listed FTP file titles. In 1993 the first search bot (WWW Wanderer) was used to generate the Wandex index. Late 1993's JumpStation was the first 'true' SE in that it used a bot for discovery, built an index, and used a web form GUI to query results.

Those early SE result lists returned simply a file name or URL - hopefully the naming convention was descriptive, often it was not. Eventually, title, meta description, and/or meta keyword were also retrieved and returned or used to assist in determining relevance. Much simpler but still basically the format today.

Search engines acted as agents, middlemen, between their users, the human searchers, and the information held by the Internet/WWW. What they knew then was names/titles, descriptions of or terms contained, and an address - and that was what they returned as answers on SERPs (Search Engine Results Pages).

Once the SEs had mostly indexed the web, certainly far beyond their ability to return all that they had discovered to a query, so were largely on maintenance and in quality control mode the logical next step as IR agent for their users was to return direct answers where feasible rather than always provide redirection. Domains such as calculations and conversions, time and weather, were obvious low hanging fruit.

This step shocked many. And definitely damaged many a business model or hobbyist's popularity. While foreseen/forecast by a few most affected and many awakened to concern webdevs cried foul. They had (and many still have) the relationship between themselves and search engines backwards - they incorrectly believe that SEs purpose is to deliver searchers to them when in reality, from the very beginning, the search engines purpose is, as it always has been, to deliver answers to the searchers' queries. All that had changed was the addition of a SE answer capability.

<aside>
One sees the same ackbasswards in regards to AdSense. AdWords (G Ads) was created to monetise Google's SERPs. AdSense was simply a way to increase revenue by monetising the longtail, those searchers who left the SERPs by clicking a result link. The purpose was to extend AdWords beyond Google, Google was acting as an agent of the advertisers not the publishers. In both SERPs and AdSense the publisher is the fodder, the product not the client. Silly publisher!
</aside>

As time passed SEs have various other answer formats, largely following a similar methodology wherein initially the answer includes a diminished but extant link to where the answer was retrieved. This was and is not done to be nice or fair or what have you, it was and is done because (1) web derived info is full of errors and omissions, and (2) the algos start out stupid; it's legal CYA behaviour. Over the years real people have been inconvenienced, injured, even died from inaccurate search result directions/information. Once a SE believes it has sufficient domain expertise it drops the already minimal link and simply says 'here's the facts, ma'am!' courtesy of us. Not anybody else.

Facts are not copyrightable.
Fair Dealing, Fair Use are legal exceptions to copyright.
If a publisher does not understand and mitigate the above their business model is fundamentally flawed.

SEs are agents for their users, ad networks are agents for advertisers.
Publishers are the, increasingly commoditised, product of both SEs and ad networks.
If a publisher does not understand and mitigate the above their business model is fundamentally flawed.

@brotherhood of LAN: the reason you are having to grapple with Search Engine Philosophy Regarding Answers to Questions is because many/most people have a mistaken conceotion of what a search engine is and for whom it is acting.
A search engine is an IR system (and a business) and it is NOT acting on behalf of the sources of the information it retrieves. That some benefits accrue to some publishers for some time is merely a (marketing) consequence of the main intent.
It never has been else and I see no reason for this to change in the near future.

tangor

6:05 am on Oct 3, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Web for Gnat Attention Spans is the current game plan. Right or Wrong we will see more of it than less.

Sad part is those who code the engines will have their version of what is "information" and that will skew the results even further.

Meanwhile, that is the playing field and one does their best in that regard. :)

And can only hope that the coders (who create the AI stuff) are fully baked, have no bias, and are really interested in providing the BEST answer for the instant gratification crowd!

brotherhood of LAN

8:31 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



<bump>

Loads of people here have been around long enough to see the evolution of search, and long enough to form an opinion on how people should find information on the web via search engines/portals, that's why I posted here.

Feel free to throw in your thoughts, random or otherwise.

Selen

8:41 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



Random thought: it is a completely false assumption that the answer / knowledge boxes only provide "simple answers."

Most of the answers give the "meat" of the content (that includes pictures, videos and other form of media). Some answers have so much content that they have a scroll like a regular web page. Others have their own tab menus to give more details.

brotherhood of LAN

8:55 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Random thought: it is a completely false assumption that the answer / knowledge boxes only provide "simple answers."


If I need to define simple vs complex, anything that a search engine can take from structured data would be simple, anything 'guessed' from wild web content would be out of context.

In that sense, I refer to the 'simple' answers, as 'complex' involves arbitrary search engines and guesswork.

Selen

9:00 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



In essence, structured data = HTML, so full web pages can be considered "simple answers" too.

brotherhood of LAN

9:05 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



No...

basically major search engines used structured data from wikidata, like dmoz in the past.

Some search engines like Google extend that further and do guesswork on HTML. Most structured data in HTML5 is in schema markup or some educated guess on the search engine's part on the nature of content.

It's besides the point, I'm talking about whether search engines, as you see how a search engine should operate, should serve structured data beyond 10 blue links. i.e. are they just a gateway to the web or something more.

Selen

9:18 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



If we consider a search engine = Burger King.

You go to a Burger King and they offer you free samples of food: from chicken nuggets and grilled ribs to ice cream and cola. You taste 10 foods and you got lunch for free. You repeat it every day.

So if a business model is to give free food, then it may work. But most webmasters cannot work with this concept in the long term.

brotherhood of LAN

9:20 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Analogies, no need! I'm asking for opinion.

Selen

9:26 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



Feel free to throw in your thoughts, random or otherwise.

Analogies, no need!

wow, ok ;)

brotherhood of LAN

9:30 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Why would we need analogies? It's a search engine, the web, people using it, people creating content and your idea of how they should work together. They're pretty fundamental.

Sorry if it sounds blunt, HTML isn't about the question.

Selen

9:34 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



Because you, willingly or not, asked for them. An analogy is a part of a "thought." They are also used to express an opinion, like it was in my case.

brotherhood of LAN

9:39 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Oh, ok no offence intended and you did caveat it with random thought. I don't get the burger analogy though. I'm just trying to approach the question of additional data beyond the '10 blue links'.

Selen

9:45 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



The answer is simple - for search engine users it's always better / more convenient to get content right on the search results page without having to visit another site. For content creators - it's the opposite. As time goes by, search engines will have to provide more and more answer boxes to satisfy the user; if they don't, users will seek other places. The genie is out of the bottle and there's no coming back to just 10 blue external links.

brotherhood of LAN

9:52 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Fair point, but do you agree that's the right way forward for all involved?

Selen

9:55 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



No, it's not the right way (the right way would be 10 blue links only).

brotherhood of LAN

9:56 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Because you believe a search engine should merely be a portal to the rest of the world wide web?

Selen

9:58 pm on Oct 5, 2019 (gmt 0)

10+ Year Member Top Contributors Of The Month



Exactly..

tangor

10:09 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



^^^^ I'm with that!

brotherhood of LAN

10:19 pm on Oct 5, 2019 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



Learnt a lot. Glad I asked. Thanks.