Big Shuffle in the All Search. Will it last?

SLocialism?

And that has never been gamed? Come on. In 90% of the cases, there is just one or two words slipped in that has nothing to do with the place, but will generate more traffic. And that, to me, is more annoying because it is really false advertisement.

Right, because no one will use the same keyword OVER AND OVER AND OVER AND OVER again, or put in keywords that don't fit...

There is no such thing as a system that is perfect. A truly random system can deliver crap for quality and low relevance. A highly relevant system has people that actually care about the keywords and filters out crap, although people will compete heavily to get to the top. Which is worse? I'll take the filter.

There lies the problem....the complainers don't understand the consumer, nor do they understand how ALL search works. It's a waste of time explaining it on a forum. You're better off doing what i do...i get approached by some content creators requiring help with their rankings, so i take time out explaining how it all works in principle without giving away what i consider the optimal settings are at the time (and LL can adjust those anyway). By telling them what to focus on, i have had them come back at a later date thanking me for improving their visibility.

I'd like to thank Phil for all the useful info provided both here and in-world. He certainly knows about Optimisation and how the GSA works more than most people here, especially the critics who just whine away on RA talking smack!

Also like to thank you too Marcel...our sharing of knowledge has also been useful

Which explains why Traffic with ALL search carries little weight....as the company position in no.1 spot with 100k traffic will get the same 12 IBL's as the person positioned no.99 with say 7k traffic. Even being on page 2 is no great shakes you just lose 1 extra IBL...which you probably can make up by just having 1 extra Pick.

Exactly. The top traffic place gets a mere 12 IBLs for the traffic - very small potatoes. An extra 12 picks from friends will match that.

LL have invested too much time (& money) on GSA, they are not going to abandon it because of a couple purists who don't think it's fair. The old ALL Search was far worst....remember all those 16sqm mirco parcels with keyword spam.?
ALL search is about the fairest system at the moment, muchr fairer than Places Search....and the person with the most picks doesn't always necessarily get to the top keyword rankings. I have a number of top spots with just a couple of Picks and ahead of those that have 20+ picks or more....that kicks Sling's theories into touch!!

Lol too funny.....but to some like Sling and Ordinal thats how SL should be!

And there isn't an alternative - i.e. there is no square peg to fit into the round hole that is SL. The reason is that search engines are for indexing and ranking pages. Google's GSA does that extemely well, and most likely much better than any alternative there may be out there. But, like all search engines, it requires pages to index and rank and SL doesn't naturally have any pages for it. So LL created some pages for it. Unfortunately, the pages are not like most other pages. Pages are normally written/created by authors, but SL's pages are not. LL squeezes a page-free system (the round hole) into a system that requires normal run-of-the-mill pages (the square peg).

It works quite well, considering that, and there is nothing better for this round hole out there. Any off-the-shelf search systems there are out there have the same requirements as the GSA - a bunch of pages to index and rank.

Agreed...and this happens in some of the more expensive Classified adverts too.

You are attributing positions to others that they don't actually have.

Over and over in these gaming threads, people have attempted to assert that gaming techniques like traffic bots and systematic Pick buying are all simply marketing techniques no different to advertising and promotion.




It's not. The most obvious bot in a box will eventually get zapped. Bots will move to the ground, with better bot systems making them appear to be normal residents.



Nope. I'm not concerned with marketing tools. I'm concerned with search-gaming tools.


Picks have already been tarnished by systematic Pick buying systems. You earned your Picks. All around you others are systematically buying them.


You're spiralling of into fantasy there.


Now you're somewhere out beyond Ursa Minor.


Edge of the second Galaxy past the traffic lights.


About to discover if the Universe is finite.


You're right on the galactic equivalent of the edge of a continent.


You've broken through.



Oh my! It's.......... beautiful! The colours. Those huge swirls!




What you were doing in your impassioned post was attributing ridiculous propositions to "the other side". It's a very well know and very lame debating technique.

Are you suggesting that because one thing has been gamed, then it is OK to game other things?

Keyword stuffing and false keywords are very much out in the open.
If someone cares to take issue with the abuse of page text then it's very black-and-white.
If false advertising annoys you, why not agitate to have it banned?

Are you suggesting that an indexing algorithm is incapable of ignoring keyword stuffing?
Keyword stuffing has been around since shortly after the first web search engines appeared. It would be a very junior school project that produced an engine that did noting about stuffing.

In SL, the only thing that makes sense to index and rank on is the parcel text and content.
For the GSA in SL, there are no sensible IBLs.
IBLs are part of its ranking mechanismm, but SL can not provide any that are appropriate.
You're not suggesting that absense of IBLs would make GSA crash and burn are you?
If that were so, then it could never index a page that had no IBLs.
IBLs are simply one part of GSA's ranking that is not applicable here.





The appropriate web search engine within SL is that of the early days of search engines These indexed simply on page content. Add in the techniques developed to combat the gaming of the page content and you have the most appropriate web-based engine for SL parcels.
In the special case of SL, you can attempt to introduce unique ranking that relates to avatar choices and activity. However, unless you are prepared to stamp down on abuses of those factors, you're just wasting your time.

/me blows the smoke off his bot cannon and says, "The only good bot is a dead bot!"

(jk, but how can a traffic bot be "good"?)

First, most customers are apathetic to anything other than finding what they're looking for.

However, most well-informed customers, wish that rankings reflected actual popularity, rather than being an indicator of how far the shop owner is willing to go to appear popular regardless of actual popularity.

Try this question, Phil: "Would you like rankings to reflect actual popularity, or would you like them to reflect how hard the owner works just to get high rankings?"

That's a different question than, "In order to get a high ranking on the search page, I need to run bots, because everyone at the top of the ranking is running bots. Does that bother you?"

Anyway, I'm a customer of yours, and I'm against use of bots to affect rankings. So, you have now seen one. BTW, I didn't find you using search. I don't use search much, because it finds places that run bots rather than places that are popular.

Gaming of ranking sucks. The fact that many businesses do it forces other businesses to do it to compete. Anyone who thinks there's a better justification than that is rationalizing. Of course, those businesses who allow themselves to be forced are contributing to the problem, and IMHO, deserve a measure of scorn, and should simply ignore those who heap it at them rather than making foolish arguments justifying themselves.

Those who heap scorn aren't helping much either, though.

In any case, I disagree with anyone who talks about "gaming" and "keywords" in the same breath. As I see it, you pick the keywords you think people looking for your products will use. That's not "gaming", that's choosing the ideal set of keywords. That's what you SHOULD do, if you're a decent businessperson.

Perhaps there's a way to game keywords I'm not aware of. Repeating keywords shouldn't be a problem because of the limited space in parcel description.

All traffic bots are good. They improve the rankings so they must be good

YES. If it can be bought (which is what running traffic bots is, effectively) then it's not valid.

The point is, you can get the top ranking without having good products. Of course, if you fail to sell, you'll fail as a business. So, to succeed at this game, you have to *not suck*, and work real hard at gaming the stats.

Seriously, I understand why you feel you're forced to do it. But that doesn't mean that it's good, or helpful, or even harmless. It sucks, and you just choose to suck along with the rest, rather than have low rankings.

Exactly!

Agreed up to a point. I would think that most people would prefer that the rankings reflected the popularity of the product and not the false popularity of feet on the land. That's something that LL's use of traffic could never achieve.

I would like them to reflect the popularity of product - not the false popularity that traffic used to be used for.

When I asked customers what they thought about me using bots to get them there, we didn't go into lengthy discussions on it. I didn't even say why I felt it necessary to used the bots - just that I used them to get the high ranking that brought them to the place. None of the replies indicated anything negative about my use of them.

That's nice to know I am also against the use of bots for traffic rankings, and have been for a very long time.

I agree with that too. What I don't agree with is the idea that some people have as to what gaming actually is. To some, doing anything at all, with the intention of improving the rankings, is gaming. That's total nonsense, of course.

I am one who started to use traffic bots in order to compete with two things. One was a place that topped the rankings for low prim furniture but didn't actually sell any low prim furniture. The other was a place that was also up at the top for low prim furniture due to the traffic generated by a club. It had a tiny shop attached but the "popularlity" of it for low prim furniture was due to something completely different - the club. I've never made any foolish arguments to justify my use of bots. I've given reasons, and I've reasoned in the discussions - and my reasoning has always been sensible and logical.

I've said many times that I prefer that the traffic rankings didn't exist, so that there is no need to use traffic bots. Unfortunately, LL has done nothing about it yet. They've merely made things worse for users in an attempt to appease a few people - many of whom have swallowed it. Traffic bots aren't going away because of the new policy. Many will merely morph and contribute to lag, which they didn't do before.

Repeating keywords for the traffic rankings is useless. It only needs one instance to be listed and the traffic takes care of where it's listed. Repeating them for the All search does make a difference, but there's nothing wrong with it as long as the keywords apply to the place and are not false. It's nobody else's business if a place chooses to put, say, "furniture" many times in the place's name, description and objects to show in search, and nothing else. If someone doesn't like it when they see the page, don't go there. I wouldn't do it because I'd expect enough people to avoid my place when they see it, but I can't criticise anyone for doing it. I used that example because I saw a page exactly like that yesterday. There was no other word than "furniture" in the name, description and objects, and many repetitions of it. It's nowhere near the top

I've used the following example many times concerning web search engines:-

Suppose there are 100 hotels in New York, all equally relevant to the searchterm "new york hotels". A search engine has to list the results in some sort of order, but how does it decide which of the 100 hotels to put at the top and which to put at the bottom? How does it decide which will be on the first page, where everyone doing the search will see them, and which of them won't be anywhere near as visible? It's algorithms decide, of course, but they can't decide which of the hotels are the most relevant to the searcher - they *all* are. And yet some hotels languish on page 10, where nobody ever finds them, while others sit happily on the first page, taking most of the bookings. Why should a hotel on the 10th page settle for that? Why should it be content for other hotels to take most of the business while their equally relevant hotel gets none? Why shouldn't they try to make their pages fit the algorithms better?

That's the situation that a few people here would like to see. They really do want people to do little or nothing to influence the results and for the results to fall where they may. They don't want people to aim at climbing the rankings. That would be fine as long as it's simply what they would like. But when they come demonising the lower ranked hotels (equivalents) for doing some perfectly reasonable things to move their places higher in the rankings, they are way out of line - and totally thoughtless.

IIRC from talking to the guys at my rl work's corporate office they can change the way things are considered in the search for relevancy since they have it looking at server(1%),directory(2%), file name(10%), extension(10%), and words in the documents(77%)

So they could change the maths to something like this for all search:
Picks 0.5%
Traffic 0.5%
Group Related 0.5%
Items listed on parcel 2.5%
Keywords 90%
Parcel Name 5%

There's the problem.
How can you measure the popularity of a product?
The number of sales? No. That's immediately gamed by bogus sales to bots whe get the recycled money back to buy again.
The value of sales? Ditto.
Some sort of voting system? No. That's gamed as well.

Can anyone suggest a system of measuring popularity that won't be gamed into uselessness?


I wonder how extensive this "some people" is?
I have seen posts like Mickey's above that try to paint anyone who objects to gaming as also being against marketing. I can't recall posts claiming that *anything* done to affect ranking is cheating.



It is a very poor ranking algorithm that used repetition of a word as a ranking factor.
The idea that, for example, a New York hotel should rank higher than another New York hotel simply because it uses the word "hotel" more frequently is just plain silly.




Who exactly are these "few people" ?

For my own part, I'm happy to demonise both low and high ranking places that do something *unreasonable* to improve their rankings.

nobody actually cares if his furni shop is 'popular'.
they care whether or not he sells whatever it is his search results have led them to believe he sells if he has a 25000 traffic place listing for 'one prim lamp', youre supposedly looking for a one prim lamp, not the 25000 traffic.
its a shop, not a nightclub. the furni is the product, not human company.
youre all upset because there arent a couple dozens people jumping around in the shop? arent you there to buy furni? i suppose i just dont understand.

Nope.
I have no such theory that the number of Picks is the primary ranking factor. It never arose in my mind that places with higher numbers of Picks should rank higher than those with lower simply because of the number of Picks.
What I have said about Picks from the time that Pick-buying became a practice is that such a practice is dishonest and gaming the search. That is all.

If you imagine that you have seen me posting otherwise, you are simply remembering posts by other people who tried to attribute to me opinions that I do not hold.

It's called "Humor", Sling.

But I see that Picks is next on the list.

Buying Picks is just like buying an Ad. It's a billboard. If someone wants to spend their advertising dollars on picks....there's nothing wrong with that. If a person wants to make a few lindens by posting someone's ad in their picks....that's their right to do so.

It's just an Ad.

"It's called "Humor", Sling. "
- and I responded in kind.


Picks -buying isn't "next". It's always been up there on equal terms with traffic manipulation.

If you think that buying Picks is just like an ad, then what do you think would happen if LL removed Picks as a Search weighting factor? Would people still pay others to have "Ads" in their Profiles?
How many Picks systems existed before LL created their new All Search and put Picks as a ranking factor into it?

Buying Picks is not Advertising. It's gaming Search. The only rationale for making Picks a weighting factor would be that they are *not* paid for.
If the intention of the search system was that ranking could be paid for, then why not just sell Search ranking in the same way that Classified are paid for?