Marketplace Search Beta Now Available

[Reid Parkin]

---

CommerceTeam Linden wrote:

 While it's clear even from discussions in just this thread that it's impossible to satisfy everyone all the time, we do this beta process and ask for your time and feedback precisely because we want to avoid the doomsday scenarios colorfully presented here.  

---

It's hard to imagine who you'll be satisfying. I did a general search for 'plants' - so maybe you'll be pleasing those merchants with items that hardly ever sell, which are now coming up on the first pages. Old overpriced palm trees by the same merchant, one after the other - these are the top selling plants? When I use the current search, the results make sense. I didn't go any further than this one test, and it failed. Please don't release it, it just doesn't work. I'm not filling out a jira on something that isn't even close to working.

[HunniHope]

ok my tuppence on this Beta:

 

1. Fix "Relevance". This should be something that makes sense, which at the moment it doesnt, or at least explain to us how you come to the Relevance Results as they currently stand
2. Remove the Auto fill search options. I don't want to see it pre-filled with whatever title somone has given their listing, I just want to search for words i choose.
3. If clamping down on Keyword spamming then dont allow the spamming to take over in the Listing Title. Cut it down by half and give us 64 not 120 characters
4. Instead of giving us the ability to hide DEMO's from searches just remove them altogether, make them a hidden object like we do with vendor systems inworld. A Demo shouldnt need a listing as long as it is linked to correctly in the main item. I already see too many DEMO listings where clicking the "buy main" takes you back to main MP front page because the master listing has been removed.
5. Allow us as merchants to set how we wish our store page to be displayed by default, or give us a ranking system where we can set the top 12 products to be shown.
6. Fix the reveiw system. Or wipe it and start afresh with a whole new way of doing it. Ive had negative reviews, and following up on them they usually down to consumer error, or bad settings, genuine faults i fix as priority, i've even so much as had blackmail attempts when in IM about them. I dont want to sit there and flag bad reviews for removal, i just want reviews to be more "realistic" and genuine.
7. I'm still thinking on this one .. time for coffee

[ChinRey]

---

CommerceTeam Linden wrote:

First, please rest assured that we're closely following the feedback in this forum.

---

 

---

CommerceTeam Linden wrote:

While it's clear even from discussions in just this thread that it's impossible to satisfy everyone all the time...

---

That is correct of course. Every buyer wants exactly the item they are looking for to turn up at the top of the list, no matter how vague their search phrase is, and every merchant wants their listings on top.

But this change, as it is now, isn't going to satisfy anybody. The buyers certainly don't get better advice and as for the merchants, even those who are lucky in the ranking position lottery will be unhappy by the way it messes up their storefront and delists some of their items.

 

---

CommerceTeam Linden wrote:

We haven't asked you to test specific things precisely because we need generalized feedback on various aspects of your experience with search both as merchants and as shoppers.

---

 Very well, here is some generalized feedback then:

As Pamela (I think it was) already said, the search algorithm is really not the right thing to start with if you want to improve the search results. Here are four smaller projects that should be relatively easy to implement and should produce noticeable improvements fairly quickly:

 

Make it easier for customers to rate the products

Customer feedback should be an important aspect in the ranking but it only makes sense if there is enough data and right now there isn't. So:

1. When somebody logs on to MP, get the software to generate a list of purchases they've made but not rated.
2. Add a link to that page, saying something like "You have xxx unrated purchases". This link should be on every MP pages, clearly visible but not too obtrusive. At the top right corner, right next to the user name may be a good place for it.
3. Make reviews optional, not mandatory. The threshold for leaving a quick star rating is so much lower than for writing a fairly detailed review and most buyers should be perfectly happy to do one but not the other. The software is already able to handle rating without reviews, all you need to do is reenable that option.

 

Exclude merchant names from general search

Well, obviously. Is that really something you need to be told?

 

Remove expired listings from the search results

That isn't going to be much of an improvement but it will be enough to noticeable, it is one of those rare changes that actually should satisfy everyone and it should be fairly easy to implement. The software is already able to identify expired listings, they're marked as such on the item pages. Why not do that check before the search results list is created? (That's not a rhetorical question btw., I don't demand or expect an answer in any way but it would actually be interesting to know.)

 

Improve the way you're processing flagged items

This is actually to make your job easier but it should benefit all buyers and all honest merchants.

When somebody flags a listing for keyword spamming, add a field to the form where they can enter the keyword(s) they reacted upon. Maybe a field where they can write a short comment too.

When somebody flags an item for relisting, add a field where they can enter the original listing number. Most probably won't but some will.

Those new fields will have to be optional, not mandatory of course and it's up to you to decide whether the added data is helpful in each particular case. But I promise you, it will make your job so much easier and, relevant to this topic, it will help reduce the number irrelevant search results.

 

---

 

Once those low hanging fruits have been taken care of, it's time to look at how search results are ranked:

• Keyword placement and density are out, let's just agree on that, OK? They have no relevancy whatsoever to any search engine of any kind.
• Sales figures are not an ideal indicator for many reasons. For a start they can't be linked to specific keywords. Let's say I sell a lot to people searching for "black roses". How relevant is that for people who search for roses in general? Even so, we just can't get away from it. The poor search engine has so little ranking data to work with, it'll have to use what it has for all it's worth. It may be necessary to re-evaluate its significance and how sales are weighed but please, keep it as part of the ranking.
• Customer ratings can be very valuable if and only if there is enough data available to calculate a meaningful average/median value. The problem with linking the data to specific keywords is the same as with sales figures though.
• Customer ratings across the merchant's entire product range. Now, that's a thought. Not sure if it's a good one but may be worth looking at.
• Click-through ratio. How often does a link of a search results list spark enough interest the potential buyer actually checks it out? Easy to implement. Can be linked to specific keywords but doesn't have to be. Unfortunately also very easy to manipulate. It may be a good idea to include that factor in the ranking but it shouldn't weight very heavily.

Of course, there are probably far more sophisticated methods than these but I'm not sure if they're worth it and the Marketplace is actually a particularly tricky case for any standard search ranking algorithm so they may not be of much help.

I think search results with exactly the same ranking should be sorted in a random and continuously changing order. If the search engine has no idea what is the most relevant one, it shouldn't try to tell us, But that means storefront listings have to be handled separately from general search. We can't have anything random there. And of course, make sure ranking order doesn't change when somebody goes to the next search results page. That's just silly.

 

---

 

Next up is what you can search for:

• Boolean search: mandatory
• Exact matches for multiple word search strings: mandatory
• Fuzzy matches: Not sure

 

---

 

But the big problem here is still the sheer number of listings. You say the new engine will be more able to deal with large volumes. I suppose that means it can process the data faster. That's nice  but it won't help because the real problem is that the poor customer will have to shift through all that data too and sadly, customers can't be upgraded. I'm afraid we're stuck with the same old model there.

More and better filtering options will help. Here are some options:

• Land Impact. In a way that already exists but it's done wrong so it has little value
• Render weight. Should be very relevant right now with Quick Graphics and all that jazz
• Original or derivate work. Is this something the seller actually made themselves or did they just texture a full perm mesh? Is it a gacha prize or some other item they bought and want to sell on? Is it something they found on the internet or in a computer game and just uploaded to SL? This can be a bit tricky since quite often there is no clear line between original and derivate works.

 

You still have to deal with that volume of listings though and of course, you can't delist or refuse to list items just because you've run out of space. There are still quite a lot you can do though:

• Expired items. Sold out? Delist. Automatically. Only make sure they're relisted automatically too if/when the seller resupplies.
• The Real Estate category. Should be moved not from the Marketplace but from the general search. Make a separate database with search and listing functions specially adapted to this rather special category. It won't reduce the volume very much but every little bit helps and this will have so many other positive effects too.
• Demos. Include in main listing. I know, that is a very difficult task and that is why you probably shouldn't try to handle it right away. Sooner or later you will have to deal with it though and it's not going to get any easier. Edit: while I was writing this, Sassy very sneakily posted a fairly easy way to fix that issue - at least temporarily. (And to make it even mroe embarrasing for me, I ahd already weritten but not psoted how a similar method could be sued for expired items - didn't occur to me it was appliable to demos too )
• Color variants and other versions of the same product. Same as demos but probably even trickier. Needs to be done though.
• TOS violating listings. You need to find ways to handle flagged items more efficiently. That is why I suggested some changes to the flagging form at the beginning or this post. But you also need to process them smoother and more precise. That autofill bug for a start. Fix it! Now! And when you come across listings that are accidentally keyword spamming because of that bug (they should be easy to spot), don't bother the poor innocent merchant with it. Just remove the offending keywords and say nothing more about it. If somebody is caught keyword spamming, send them a polite message telling them their item has been temporarily (that's an important word here) delisted and ask them to fix it or tell you why they disagree with the decision. Also ask them to check their other listings to see if they have similar problems. Give them a time limit and then recheck. And, most important, don't forget to tell them which keyword it's all about! Repeating offenders should be removed and even temporarily or permanently banned from selling on MP but even they has to be given the chance to explain and argue their case first.  We want the merchants who deliberately and consistently abuses the TOS off the MP but we don't want people to be punished that severely just because they've been thoughtless or ignorant.
• IP violating listings. Now, that's a minefield. I definitely don't envy you that job. Fortunately for you, that is something Linden Lab as a whole will have to deal with. For now I suppose you will just have to do the best that you can with the tools and means available.

 

---

 

This ended up as a long post but this is what you actually need to do. And if I can be a bit blunt (again), very little of this should be news to you!

The new beta search doesn't seem to address any of the MP search related problems for anybody in any meaningful way. That is why I suggested you should just scrap it and start anew. But if you want to build on this new one, then OK, do it that way.

But please, next time you present a beta version, make sure you have something to show. It doesn't have to be much, but it has to be something.

[Reid Parkin]

---

HunniHope wrote:

1. Fix the reveiw system. Or wipe it and start afresh with a whole new way of doing it

---

I've seen this come up before. To fix it, yes. To just wipe away reviews, no. People take pride in the genuine good reviews that they have accumulated over time. It would be unfair to just wipe it. Tke a look at a store like Dutchie Furniture - do you really think all these reviews should be wiped? How about Studio Skye, just get rid of all his reviews as well, huh? Stop just seeing it through your own eyes - not just to you. This silly idea was brought up before. 'Get rid of old reviews' or something like that.

[Polymath Snuggler]

YAY a reply!  ♥

 

Okay !  Cynicism, dark humor and irritability aside:  Here's a long winded in-depth analysis for you Commerce Linden people!::

 

The current Marketplace Sales Environment (non BETA) suffers from some serious but non-critical flaws.

Here they are in rough order of priority.

 

1: Dependency upon customer reviews:  Reviews are an IMPORTANT part of a virtual marketplace economy.  (see Amazon.com, Yelp! Newegg etc etc )   The problem with the current iteration of SL Marketplace is that customers rarely if ever have any incentive to place a review for an item unless they're dissatisfied with it.  I've sold roughtly 1500 units in the past few months and the number of positive reviews I've gotten without interacting with my customers has been 2~  I think?  That being said, I have people who bubble and cheer effervescently in IM's with me over my work.  But they don't stop to think "hey, I should leave this girl a good review" unless I specifically nudge them towards that direction.

Properly solving this problem unfortunately involves developing new client-to-website code I believe, as the most assured way to have people review things they've purchased would be to have an inworld "rate me" popup ( typing optional ) after an item is delivered ( or maybe removed from the customers "recieved items" folder? )    I understand in-world to website integration is hitchy and difficult to implement at best.  But this would fix most of the reviews problems as "revenge" reviews would be drowned out by the many valid ones.  Likewise obvious review exploiters would be less impactful overall.

 

2: Inclusion of Content Creator and Store names in the general search field:  This singlehandedly muddles the marketplace up for certain queries to the point of utter unusability.  There's even a "Merchant / Store" alternate search field right at the top of the page for people to use that SHOULD DO THIS. But not the "General Search"!!  It's like the idea was thought about, but never implemented!!  Just to use Trompe Loeil as an example.  ( Cory is an excellent SL merchant, she's a great example of what we aspire to )  But if I'm looking for Edo era Japanese paraphanalia, chances are I'm probably not looking for Trompe Loeil furniture.  The problem is I can't find anything relevant to my actual search because I'm having to sort through her entire catalog of items before I find anything else.  Here's where it's important to distinguish:  Cory deserves every single bit of her earned search ranking.  She deserves every single bit of her visibility from sales popularity.  When I search "White Bed" she should have the #1 result ( actual return is like ~ #4 on the current system !!  working as intended! ) Just not when I search for "Edo".

 

3: Lack of priority for newly posted items:  Posting a new entry on Marketplace should provide some measure of visibility for a short duration of time.  ( 7-14 days maybe?  though I'm sure some people would find a way to exploit this )  It shouldn't outshine the tried and true best-sellers, but having your new products place in dead-last is always a tad bit disheartening.  Perhaps giving stores the option to customize their own storefront so that their shop-main page can have a "New Releases" bar in the same spot where the "Featured Items" bar exists for the standard listing results would be a nice addition to the MP experience.

 

4: Keywords: Keyword spam / abuse.  Keywords & Catagories used to be absolutely essential for a proper search, back in like 2007.  The problem was in the exploitability of this.  For a few years google tried various tactics to prevent abuse of keyword & tag spam but they ultimately gave up on trying to prevent exploitation of it, and opted to just not use them at all.  Instead they scan the text of descriptions and actual website contents for parameters that they were looking for and used that in lieu of the tags.  Since the SL Marketplace was designed around the keyword system though, I don't think removing them entirely would be possible as everyone would then have to redo every listing they own.  If we learned anything from the VMM implementation, some people have literally 9500+ listings...  I'm sure people will put forward recommendations on how to solve this issue in this thread.  But this is probably the most difficult problem to solve.  The best 'solution' may be to just leave it as is.

 

5: DEMO items: Demo clutter is not helping things.  DEMO items should not be searchable.    When listing a product you should be able to hit a "radio button" that sets the product as a DEMO, and excludes it from search listings.  Upon implementation, retroactively any item with the word DEMO in the title should be dumped into that catagory ( pending shop owner approval of course ).  In addition, currently only being able to link one item to a given "DEMO" listing is a tad bit crippling.  Multiple listings should be able to be tied to a single "DEMO" item for organizations sake.

 

-------------------------------------------------------------------------------------

That's all I can think of at the moment.   A lot of this is paraphrased from other parts of this thread~ as well as feedback from people inworld etc etc.

 

Things the MP currently does RIGHT.

 

Listing things by sales popularity is key to the success of any online marketplace.  It HAS to be.

 

I've noticed what I percieved as a "store popularity" statistic as well ( sorry if that's some sort of secret part of the algorithm )  where new items from a popular store are more likely to fly up in the visibility rankings and stay visible based on a store's overall weekly / monthly sales than new items from an unheard of store that doesn't sell much of anything.  While this sort of mechanic makes it difficult for new stores to gain a customer base #3 in the above essay may help alleviate that minorly.    I believe having this mechanic is important though as top-end content creators are likely to continue being top end content creators.

 

Thank you LL for taking the time to consider all of this.   I think fixing your marketplace code will help both Second Life and Sansar to be stable profitable enterprises for content creators and Linden Labs alike.  I seriously would not be where I am today without SL as a platform to create and market on.

[Darrius Gothly]

My turn ...

---

CommerceTeam Linden wrote:

First, please rest assured that we're closely following the feedback in this forum.  You're not spilling ink in vain.  We haven't asked you to test specific things precisely because we need generalized feedback on various aspects of your experience with search both as merchants and as shoppers. 

Based on the feedback you have given, we're working on making incremental changes.  We'll promptly notify you of each new deploy (and yes, we'll list the changes made) and ask for additional responses.  While it's clear even from discussions in just this thread that it's impossible to satisfy everyone all the time, we do this beta process and ask for your time and feedback precisely because we want to avoid the doomsday scenarios colorfully presented here.  

---

I'm going to start waaaaaaaay down at the bottom level in the hopes that you not only cherry-pick the technical feedback provided, but also in hopes you pick some of the basic business ethics and practices discussed as well.

1. The forums are a communication tool. They are here not only for Residents and Merchants to communicate, but also for Linden Lab employees to communicate. When you say "rest assured" .. Rest assured on what? Past experience? Commonly held expectations? Those two are at opposite ends of the scale, and if we "rest assured" on past experience ... we're resting on the realization that no one gives a hoot. Give us some positive past experience to "rest assured" on and then you can "rest assured" we'll be more receptive and open in the future.

2. Drop this stupid "Commerce Team Linden" moniker and own your job. Faceless titles posting in the forums have no weight, no authority and no "skin in the game". Every one of the Merchants that has posted in this thread (and many others) has massive amounts of "skin in the game". You get paid WAY more than most of us do, and you can't be bothered to step up and take responsibility for your team, your product and your duties? Then Pfffftttt! Why expect us to give you any respect or sincere consideration. Become a single point of authority that acts as the focal point and you will find us focusing on the real task, not faffing about with time-wasting goose chases.

3. Clearly state the goals of the task at hand. You've paid someone good money to create new code. In today's market, and even hiring total newbs to do the coding, you've invested a bunch of your employer's money. So invest a bit more and do everything you can to ensure maximum return on the money already spent. Tell us what changed, tell us what needs looking at, and be clear what needs to be discovered through "real use". The generic "it's new and improved" doesn't tell us anything, sets false expectations that the 100's of suggestions already posted and explained fully will be implemented ... and sets the whole exercise up for inevitable failure. If you don't explain what needs doing .. NOTHING USEFUL WILL GET DONE! (Management 101)

4. "Working on" and "incremental changes"? Okay, obviously you have not actually read what's been said. The "incremental changes" that need doing are clearly restated. Fix Demos. Fix spamming. Fix Relevance.

===========

The feature (Marketplace Search) is one of the most critical features of Marketplace. Shoppers will not "browse" through an online store with as many items as SLM. They will almost always SEARCH for what they want. You've got the web stats. Go look at them and tell me where visitors go when first logging in to SLM. Search!

This is no big secret. As evidenced by the vigor and energy contained within the words posted on just this thread alone, you can clearly see it is a feature that is crucial to every single Merchant using the site. So don't go shaking the tree like you have without recognizing how tender and important this is.

You do NOT go into a store selling fine breakables then start playing catch with the merchandise! Likewise you do not tear up and rebuild a feature as delicate and important and EXPENSIVE as Search without clearly expressed and visible caution. You've grabbed the "tenders" for most of us, now act like you understand that! Handle with care .. dagnabbit!

[Pamela Galli]

I agree with your points, Chin, except one that I strongly disagree with.

One of the positive changes from Xstreet was that customers must give information, and not just stars, in their reviews. The reason this is essential has to do with the purpose of a review, which is simply to provide potential buyers with information about whether the product is as described in the listing, the quality of the product, and the quality of customer service.

 

Stars alone provide no useful information except maybe 5 stars, which may be assumed to confirm that all these are acceptable. A one star review can mean anything -- and in my experience most often has nothing to do with the purpose of a review but with the buyer not reading the listing.

For example, does one of my houses, the listing for which clearly states in all caps, not once but twice, that no furniture is included other than that listed, deserve one star?  Or, I sell both dining sets and individual tables alone -- the latter clearly stated in all caps -- should the table get one star because it is not a set?

In other words, should a product get one star because it is exactly as described in a listing the buyer didn't bother to read? And if not, how else would this be clear except in a written review?

And above all, should a product get one star because it doesn't fit an 8 ft avatar, and the buyer did not try the demo provided on one of the four sims I pay for in order to provide a demo?

Informative reviews, the ones that provide useful information,  are one of the things the MP actually does well, tho they could make it easier to leave them. Stars alone can be used simply to to punish or extort.

[Pamela Galli]

To add to what Darrius has said:

The marketplace Is treated like some unimportant backwater of SL, something LL employees are assigned tasks regarding, when they have spare time. Dakota is the only person I know of who takes public responsibility for MP related matters (and does a good job of it). 

I would like to see some evidence of a dedicated MP team that takes ownership of the MP as a product, as I take ownership of my products. The last time I knew who was on the team (Brodsky and Grant), they were all fired. Since then, I see no evidence there is a dedicated team.

 

[Darrius Gothly]

Dakota! Now there is a name I've not heard in a long time. Yup, I'm absolutely gonna agree with Pamela on this point. Dakota was one of those that understood how to listen, how to respond .. and how to provide good customer service. Of all the things that can be click-drag-copied in SL .. Dakota is one I'd copy..copy..copy..copy..copy all day long!

[ChinRey]

---

Darrius Gothly wrote:

Dakota! Now there is a name I've not heard in a long time. Yup, I'm absolutely gonna agree with Pamela on this point. Dakota was one of those that understood how to listen, how to respond .. and how to provide good customer service. Of all the things that can be click-drag-copied in SL .. Dakota is one I'd copy..copy

..copy

..copy

..copy all day long!

---

He's still here I hope! It's not that long ago he helped me with an MP issue and yes, his customer service was absolutely spot on!

[ChinRey]

---

Pamela Galli wrote:

One of the positive changes from Xstreet was that customers must give information, and not just stars, in their reviews. The reason this is essential has to do with the purpose of a review, which is simply to provide potential buyers with information about

whether the product is as described in the listing, the quality of the product, and the quality of customer service

.

---

I see your point and even agree with it. But we're talking about using customer ratings as part of the search results ranking and for that we need volumes. A single star rating, good or bad, doesn't really mean anything in that context. Even ten ratings are barely enough, at least if we're talking about medians rather than averages.

In that light, well if somebody gives you a poor rating because they couldn't be bothered to read the description, I'm sure there'll be fifty customers who give the same house a fair rating. It should not simply be a matter of finding the mathematical average or median value btw. There are ways to weigh the data and compensate for misleading input. The simplest one I mentioned earlier: ignore highest and lowest figure. There are other methods too of course but in the end the one incorrect value should simply drown in the large number of relevant ones.

Mind you, I'm not suggesting that the feedback option should be removed, far from it. I just don't think it should be mandatory for people who want to rate the product.

Come to think of it, maybe the feedback function should even be expanded into a small (optional of course) customer survey? That's a completely different project from the one we're discussing here though.

[Anaiya Ahren]

Only for this project.  The "exclude keywords" project is clearly targeted at pleasing the kind of people who have things like "D*E*M*O*"  but not the actual word "demo" in the keyword field of all their demo listings.

[ChinRey]

---

Pamela Galli wrote:

I would like to see some evidence of a dedicated MP team that takes ownership of the MP as a product, as I take ownership of my products. The last time I knew who was on the team (Brodsky and Grant), they were all fired. Since then, I see no evidence there is a dedicated team.

---

That is something I never considered before. How many people are in the Commerce Team? I've always imagined it as a group of five or six or so. If there's just one or two working full time with MP, at least we have a very obvious explanation to all the problems: they're critically understaffed.

[ChinRey]

---

HunniHope wrote:

Remove the Auto fill search options.

---

Oh, we forgot that one! Thank you, HunniHope!

Yes that one has to go. If it stays, we'll end up with a competition who can put the most asterisks at the start of their product titles, because those are the only ones that will have time to load.

A good autofill function involves a little bit more than an alphabetical listing of a quick database lookup and cool as it looks, it's not really that useful and definitely not something you should spend time on effort on right now.

Besides, do you know how much lag it adds?

 

But now, Darrius mentioned something about what to spend time on and what not to spend time on. That long post I wrote earlier today alone too almost four hours and I still haven't figured out where to send my consultant fee invoices to LL.

So goodbye for now everybody! See you again soon in another discussion!

[Pamela Galli]

---

ChinRey wrote:

---

Pamela Galli wrote:

One of the positive changes from Xstreet was that customers must give information, and not just stars, in their reviews. The reason this is essential has to do with the purpose of a review, which is simply to provide potential buyers with information about

whether the product is as described in the listing, the quality of the product, and the quality of customer service

.

---

I see your point and even agree with it. But we're talking about using customer ratings as part of the search results ranking and for that we need volumes. A single star rating, good or bad, doesn't really mean anything in that context. Even ten ratings are barely enough, at least if we're talking about medians rather than averages.

In that light, well if somebody gives you a poor rating because they couldn't be bothered to read the description, I'm sure there'll be fifty customers who give the same house a fair rating.

It should not simply be a matter of finding the mathematical average or median value btw. There are ways to weigh the data and compensate for misleading input. The simplest one I mentioned earlier: ignore highest and lowest figure. There are other methods too of course but in the end the one incorrect value should simply drown in the large number of relevant ones.

Mind you, I'm not suggesting that the feedback option should be removed, far from it. I just don't think it should be mandatory for people who want to rate the product.

Come to think of it, maybe the feedback function should even be expanded into a small (optional of course) customer survey? That's a completely different project from the one we're discussing here though.

---

This is part of the problem, which you address above: there are very very few reviews, because LL does not make it easy like Xstreet does. Only the most motivated leave reviews, and a high percentage of those are angry about some thing, because they did not read the listing, did not read instructuins, did not contact the seller, etc. Of course some negative reviews are deserved, but some are leaving reviews for things that have nothing to do with the product, but their own failure to read -- and with a star only rating, not only does no one know what it is for, but it might be the only review left.

[Anaiya Ahren]

I am completely flaberghasted over this little "exclude keyword" project of yours.

 

This is the work around hack that we were using because you wouldn't do as Sassy has rightly told you should have been done at the outset.  Segregate the demo (using an actually reliable data field like the "link as demo" field rather than the lying liers keylies of lies field).  Of course that was a long time ago and things have moved on from there.  Excluding "demo" stopped being sufficient when enough spamming cheaters determined to force people to look at their demos caught onto the trick and started using words like "dem0" in their keyword field instead.

 

That battle has long since been lost.  The little keyword exclusion project you're wasting time and resources on, is already relatively useless as a means of filtering out demos.  Any merchant who wants to force their demo into search results can easily defeat your "solution" by not putting the word "demo" in the keyword field. 

 

There's all this stuff that urgently needs to be done before the marketplace becomes completely unusuable, and you're mucking about throwing away resources to not solve a problem that should never have existed?

 

You have produced a predictable mess that you were warned about and you still won't listen to the people who warned you.  Segregate using "link as demo" field and allow demos to be linked to multiple products.  Stop mucking around and do it already.

[Lexbot Sinister]

---

"

We are still working on an important feature, which will allow you to exclude keywords (such as “demo”) from your search - but want to get this into your hands now and get your feedback! 

"

We shouldn't have to exclude demos from search results - EVER!  If we have to rely on that sort of mechanism, just like keywords it will always be subject to abuse.  Just giving us boolean NOT will not work.  DEMO, D*E*M*O, D-Emo etc.  let alone different languages.

---

ALL demos should be a separate object type! There is NO neeed to include the demo's the the search, ever. Being able to exclude keywors- are you kidding me? Like, really, this is what you point out as a new feature? I'm feeling insulted, because this is like pointing out we are straight up stupid. The NOT operator has been around for a long time already, and working on the marketplace as well.

Need i mention that if a keywords exclusion targets one specific keyword, the keyword changes? We are able to search for "whatever NOT demo" and have been able to for a long time. Why do you figure there are so many things named "d*e*m*o and similar? And if you try to fuzzy-catch all the demo iterations, it will just change to "tester" and "tryout" and what not.

The only sollution is to have it as a different object type, that can be excluded from search with a tick of a box. And flag the demos of a wrong object type.

 

The whole " we will give you a feature you already have & it will never work, because it can't " is the kind of stuff that makes my blood boil.

 

 

(Here's an example -looking for a "dress NOT demo", in the 0-10 ls span, sorting lowest to highest price to have any demo's float up. The Demo's present are the one that are dodging the NOT operator on purpose.)

https://marketplace.secondlife.com/products/search?utf8=%E2%9C%93&search[layout]=gallery&search[category_id]=38&search[sort]=price_asc&search[per_page]=96&search[keywords]=dress+NOT+demo&search[price_low]=0&search[price_high]=10&search[prim_count_low]=&search[prim_count_high]=&search[copy_permission]=0&search[modify_permission]=0&search[transfer_permission]=0

[Tari Landar]

You want my feedback, so here it is...

As a merchant, and a consumer...I don't like it, for all the reasons everyone else already said.

I recall a thread asking us for our input as merchants on what we'd like to change about MP, things we'd like to see different, things that are broken..yadda, yadda

I suggest turning to that thread(I assume it still exists, somewhere, admittedly, I didn't and won't be, look/looking for it myself). Most of the feedback in this thread, is the same feedback you got back then. Why? Because no one can tell what has or has not been changed, with these new changes. Were changes made at all? Probably, but if even the most informed and intelligent folks, who study the MP quite frequently, cannot find the changes made, it stands to reason the rest of us with less experience looking at MP so technically, won't find them either.

WHen people spend years telling you what's broken and needs fixed, and you've still yet to implement any of it, there's a problem. Of course folks are going to get up in arms, and likely even angry, about this. What is the point in asking for feedback, especially multiple times, when the feedback given falls on deaf ears(no offense meant to those with deaf ears,, I'm at a lack for better terminology, I apologize).

You say you're listening, but LL always says that. As far as MP is concerned, the evidence thus far proves otherwise.

Implement the changes as outline by the far more intelligent than I folks, and then ask us to "test" it. At least we'll have something to test and then we can give proper feedback. You can't really give "new feedback" on something that isn't new  

[Pamela Galli]

I could not find the thread Tari referered, but for a fascinating trip down memory lane:

 

A schedule of upcoming work the Commerce Team will be doing:

 

https://community.secondlife.com/t5/Merchants/Marketplace-Update-March-2011/m-p/740185/highlight/true#M799

And don't forget the Search Survey we completed:

 

https://community.secondlife.com/t5/Featured-News/Help-Improve-Marketplace-Search-with-Your-Opinions/ba-p/2384033

 

https://community.secondlife.com/t5/Merchants/At-long-last-Marketplace-Search-Improvements/m-p/2813450/highlight/true#M51105

 

 

[Tari Landar]

The thread very well may be gone now, I just remember reading it, mostly for entertainment value, knowing that none of the ideas suggested were likely to be implemented, despite being *really good* ideas.  

[Pamela Galli]

---

Tari Landar wrote:

The thread very well may be gone now, I just remember reading it, mostly for entertainment value, knowing that none of the ideas suggested were likely to be implemented, despite being *really good* ideas.  

---

It wasn't that long ago, but we carefully and completely documented all problems, and were assured we were heard. I think Ebbe even posted in it.

The End

 

[Rya Nitely]

---

ChinRey wrote:

---

Darrius Gothly wrote:

Dakota! Now there is a name I've not heard in a long time. Yup, I'm absolutely gonna agree with Pamela on this point. Dakota was one of those that understood how to listen, how to respond .. and how to provide good customer service. Of all the things that can be click-drag-copied in SL .. Dakota is one I'd copy..copy

..copy

..copy

..copy all day long!

---

He's still here I hope! It's not that long ago he helped me with an MP issue and yes, his customer service was absolutely spot on!

---

She

[Rya Nitely]

---

ChinRey wrote:

---

Pamela Galli wrote:

One of the positive changes from Xstreet was that customers must give information, and not just stars, in their reviews. The reason this is essential has to do with the purpose of a review, which is simply to provide potential buyers with information about

whether the product is as described in the listing, the quality of the product, and the quality of customer service

.

---

I see your point and even agree with it. But we're talking about using customer ratings as part of the search results ranking and for that we need volumes. A single star rating, good or bad, doesn't really mean anything in that context. Even ten ratings are barely enough, at least if we're talking about medians rather than averages.

In that light, well if somebody gives you a poor rating because they couldn't be bothered to read the description, I'm sure there'll be fifty customers who give the same house a fair rating. It should not simply be a matter of finding the mathematical average or median value btw. There are ways to weigh the data and compensate for misleading input. The simplest one I mentioned earlier: ignore highest and lowest figure. There are other methods too of course but in the end the one incorrect value should simply drown in the large number of relevant ones.

Mind you, I'm not suggesting that the feedback option should be removed, far from it. I just don't think it should be mandatory for people who want to rate the product.

Come to think of it, maybe the feedback function should even be expanded into a small (optional of course) customer survey? That's a completely different project from the one we're discussing here though.

---

I agree with Pamela here. In the Xtreet days leaving a review/feedback was optional. Although this did increase the number of ratings, it wasn't enough to negate the impact of those 1 star ratings that had no explanation. A couple of 1 star ratings has a huge impact. There were many people, including myself, who wanted the system we have now. We asked for it. We wanted to know why we got one star.

Ratings were also anonymous back in the Xtreet days, (unless you also left feedback) and this gave more freedom to dump your one star and run. With the current system of having a name with a 1 star rating at least you could IM them inworld and ask why. But how would the Commerce Team know if the rating is justified and not malicious if you flag it? I do appreciate that they can remove unfair ratings and reviews at the moment.

edited: to add the word 'feedback' to avoid confusion between review (explanation) and rating (stars)

[Polymath Snuggler]

A mandatory review should not be instated.  However a prompt to review should be attached / activated in some way after delivery is completed.  (Sassy had an EXCELLENT idea on how this could be done )  All reviewer names should be recorded along with those ratings.  Reviews should be able to be submitted without text, not everyone wants to take the time to write out a little blurb about the "why".  More often than not "3-4 stars" is more than enough information to communicate "This thing I bought sort of met my expectations."    If the shop owner is curious enough ~ they can then seek out the reviewer and ask the 'why'.

Achieving statistical relevance is important in this.  That is the goal.

 

It's worth noting that the 1 star - 5 star dichotomy is not an exclusively SL related problem.  Most review scores for other online marketplaces are predominantly 1 star and 5 star ratings averaged out to an overall 3.5-4.5 star rating.   That's just how people answer these sorts of feedback questions.

[Pamela Galli]

---

polysail wrote:

  If the shop owner is curious enough ~ they can then seek out the reviewer and ask the 'why'.

 

---

Please see my post above which includes an explanation of what and who a review is for -- and it is not the shop owner. A review is for the purpose of providing useful information to prospective buyers. A blank one star review provides no useful information other than that someone didnt like it for some reason. Of course, review information is useful to the seller as well, but the primary beneficiary is the potential buyer.

As Rya says, Xstreet had those kinds of ratings-without-reviews and I never heard anyone say he did not consider reviews with ratings a vast improvement.