Second Life’s Marketplace – A Post-Mortem

October 9, 2010 by
Filed under: Linden Lab and Second Life 

October 6th has passed and the once-dominant XStreet SL virtual goods sales website has been closed to purchases. In its stead is the new, improved, made-from-scratch Second Life Marketplace. Cue the joyous heralds, release the doves, the celebrations must begin.

So obviously it’s time for me to do what I do best, explain why it’s here and where “here” is. And to explain where “there” is … there being the destination the Marketplace is moving toward. I purposely chose the term “Post-Mortem” for this article because they have done everything they could to kill Virtual Goods sales … and succeeded beyond even their meager dreams.

The Initial Goals

When it was first envisioned, the Second Life Marketplace (or SLM as I like to call it) was intended to be a state-of-the-art shopping experience for residents of Second Life, to help guide them to the best merchandise with the greatest of ease, and thus further grease the wheels of commerce that keeps Second Life ahead of the competition. The Commerce Team voiced concerns about the then-paramount site, XStreet SL, about its age, flexibility and the ever-present impediment of year’s worth of “Clutter”. Unfortunately, in their zeal to talk up their grand plans, they also stomped all over the toes of those that made XStreet SL the success it was … the Merchant Body. In short order they totally alienated the very people that should have been behind their efforts and thus doomed the entire project to a mud-slinging battle from start to finish.

This opening posture, the “essence” of what would become SLM, was so horribly soured that they had to do something drastic to save it. Of the two choices available to them (embrace the Merchants and make it a joint effort, or close them out entirely and go it alone) they did as they have always done and locked the castle gates, developing SLM from the bottom up to contain only their vision and with no part of it to include the various needs and desires of the first level users at all. To be concise, they turned a failure into a disaster.

Starting Feature Set

The very first feature the Commerce Team trumpeted was a shopping tool every other Internet Sales site has … a Shopping Cart. Their observation was keen, clearly if all these other sites have a shopping cart then the only way we can imbue SLM with a modern look-and-feel is to also include a shopping cart. They failed to understand WHY it was present on the other sites, they just knew the “other guys” had one, so they surely must have one too.

The very next feature they envisioned was actually more a set of rules and limitations meant to enforce a uniformity that would (they hoped) result in a uniformity and beauty that could not be corrupted by any style-blind Merchant. To this end, they eliminated GIF images, shrunk the maximum image sizes, removed text enhancements, imposed limits on sizes and types of text available, and generally set very tight restrictions that served a prominent segment of the sales but worked directly in opposition to the remainder. However, in their mind, this was not only acceptable but desirable. If they remove the ability for Merchants to stretch the Style Guidelines then they remain assured that “Uniformity” will be maintained.

As a final coup de grace, they pronounced XStreet SL as the total antithesis of what SLM was meant to be. Everything that XStreet SL did was considered suspect at best and to be avoided if at all possible. In short, they wanted a site that turned the Virtual Goods world on its ear … and their method of doing so was to totally ignore the years of wisdom and purpose built-in to XStreet and invent all new paradigms out of thin air. Thus they ignored the logic behind the features of XStreet, and having dispensed with the logic, a very short step allowed them to also disregard the features themselves. Their mantra became “If it exists on XStreet then it must NOT exist on SLM.”

A Black Ops Development Project

With their starting goals in hand and their mandate to create something totally different and drastically opposite of the existing XStreet SL, they tossed a development team into a cellar, slammed the gate shut and set a deadline date. Now don’t get me wrong, there is nothing inherently wrong with doing a majority of development “in the dark.” In fact, it is the preferred method. But when you don’t have a good specification to develop toward and when you have ignored everything that made the parent product a success, developing in the dark produces a mutant beast that no one recognizes and no one feels comfortable using. It’s somewhat akin to putting on a blindfold and mixing odd random chemicals in test tubes hoping that what you produce is the cure for the common cold. Yeah, fat chance …

They did at least hold true to their initial goals. In short order they realized that in order to provide for the shopping cart and the goal of visual uniformity, they would have to drastically limit any expression of creativity. So as news began to trickle out and merchants began to get wind of what the new SLM would provide, rancor began to rise. The feedback from these first-level users was, of course, ignored completely. After all, the merchants didn’t even understand the development goals, so how could they hope to understand the requirements to bring those to reality? It does make sense … but sense in the same way that cutting off your head eliminates that awful living, breathing and moving stuff that your body demands.

The Closed Beta Has Arrived!

They toiled for months in their dark cellar, and finally one day they decided it was time to expose the site to a bit of light … so they could iron out a few tiny wrinkles that remained. Of course, they needed live data to make it work, so they secretly migrated all the listings from XStreet and stuffed them into the new SLM mold. Never mind that they succeeded in stripping the soul out of the listings, inconsequential that they totally scrambled prices and pictures and chopped the listing text … it was just test data so that’s okay.

They announced with great fanfare that they were soliciting a select group of merchants to join their closed Beta Test Team. They sent out a Non-Disclosure Agreement to those that responded so as to protect the secrecy of their efforts. Sadly that NDA also prevented any participants from saying anything about Second Life, Linden Lab, XStreet or SLM in any shape or form. But of course that didn’t matter, they had their population of willing testers … and off they went back into the cellar, new flunkies in tow.

Ahhh, but you see they failed to understand something. Those people they drafted had no intention of creating a sales site to benefit all merchants, they were firmly focused on creating a web site that served their needs only, and if possible would make life difficult if not impossible for their competitors. During the 2-week long Beta Test period, no real bugs were fixed. Instead the Beta Testers spent their time fixing their listings and reporting “Everything is GREAT! Good job.” Well gosh, when you surround yourself with “Yes Men”, the only answer you hear is “Yes”. To their credit, the development team realized their precious Closed Test Team was totally useless and thus they set off on a new course that was sure to fix the problem and result in the greatest acclaim ever raised for a project of this dimension.

A New Day, A New Beta … A Public Beta

Since their hand-picked and compliant Closed Test Team had done nothing of benefit, they reasoned that opening the Beta to every merchant would net good results. I mean, how can it fail? Let everyone see the new site, open it wide and let every single merchant contribute bug reports. This has just GOT to succeed, right?

Well … yeah … sorta. In order for it to work, you have to guarantee certain rules from the start. First and foremost you must promise the testers that no matter what happened, their sales and income would not be impacted. You also have to make it very simple for people to file bug reports and then you have to demonstrate a dedication to fixing those bugs reported. So they opened a blog thread to comments, solicited feedback, and set the hounds loose. In less than a day the hounds had ripped every square inch of meat from the bones and left nothing behind. Linden Lab responded in typical fashion by closing the thread and directing people to file JIRA bug reports instead of posting comments on the blog.

And then they set about responding to those bug reports with a disdain that spoke of great contempt. Valid concerns, serious problems, valid bugs were shrugged off, ignored or just plain ridiculed. Overall the level of response from the Development Team was striking in its diffidence and animosity. What did filter through was a contempt for what merchants wanted and a penchant for making bug reporters feel like alien invaders and not part of the “Team”. And that was just the JIRA they bothered to comment on. Most of them went unassigned or ignored, in itself a means of expressing disdain and contempt for the valuable feedback pouring into their ears. They tuned it all out, and barked back at those they actually heard. It was a disgrace.

Immovable Object, Meet Irresistible Force

Eventually the sheer volume and emotion of the responses had some effect. The paltry 2000 character limit for product description was raised to 5000 characters. It was a give, but not nearly enough. Most listing just don’t need that much and don’t use that much, but some listings do need it. Without that much room, listings are slashed and crunched as to be meaningless and useless. But what mattered most was they were protecting their vision against all attackers. The Development Team was under siege and they did their best to resist. Shamefully they never realized the “fire” they were taking was not from people wishing to harm or destroy their precious new site, it was from people begging desperately to make it usable to as many as possible without damaging the fabric in whole. With this basic misunderstanding of the intent behind the comments came the wrong response of drawing battle lines between the two sides and treating all suggestions as suspect.

(It needs to be mentioned here that the so-called “Cheerleaders” for Linden Lab did the most damage of all, wrongfully accusing those making suggestions as being Anti-Linden Lab and hell-bent on destruction. That frame of mind permeated the Development Team and thus excused them to continue their rejection of valid and needed features. In their misplaced and wrong-headed loyalty to The Linden Lab Vision, they unknowingly doomed any productive dialog to impossible at best, and unconscionable at worst.)

Even so, some features did get accepted. Eventually the Development Team did yield some ground and agreed to add some tools and features. However because these features were never considered viable or useful in the early stages, their implementation meant they had to be grafted on to the sides and gutters, ruining the clean pristine look and feel and resulting in a hodge-podge of menus and on-screen widgets that has begun to resemble a kit-bashed model and not a cohesive design. (More on this outcome later…)

Rush to the Finish Line

The Development Team finally decided enough was enough. They saw their highly speculative deadline looming and very little actual progress being made, so they once again dove in with vigor. This is that last stage of development in which the most advances are made, but is also the stage at which the most mistakes are made. From the outside, SLM still looks pretty clean, but you can bet your bottom dollar that the code under the hood is a mish-mash of unique, slap-dash and undocumented “fixes” just to meet the deadline. In short-because they were rushed to push out a half-completed product according to unrealistic deadlines, they did what they had to do … and that means writing crap code in the wee hours of consciousness, perched on the edge of mental fatigue and way past the point where the body has given up.

Opening Day – MarketPlace – The New Frankensite

So opening day has come and gone. The old XStreet SL is closed and sales are no longer being accepted there. Instead attempts to purchase from XStreet are being redirected to the corresponding products on SLM. However that isn’t always true. A fairly large number of products have no corresponding SLM listing. This is due to many reasons, but mostly it is the result of poor design choices and the permanent nature of the first “Test Data” the Development Team migrated at the start of their effort. In short, they shot themselves in the foot and have not stopped limping since.

The new features, meaning those features that existed on XStreet but were initially omitted from SLM and then later added amidst vociferous complaints … they’re there … sort of. The problem is they weren’t part of the initial design, so they have been glued on to the sides of SLM. There is no pretty way to add a lot of them, so the Development Team did the best they could do. They sandwiched new buttons into the listing pages, rearranged and shrank or grew buttons and links, and basically converted their glamorous sleek shiny website into a series of patches, hacks and workarounds. They have created a Frankensite … from birth.

The road from here is clear of course. There will never be enough time now to do it right. From here on out the only changes made to SLM will continue to be hacks and patches that yield more and more unwieldy code. They will result in a site that cannot be maintained and cannot be grown or improved. They made the wrong choices at the start, they tried hard to make their choices stick to the end, but realized they had to give some … and in giving committed the sin of creating a bastard product.

The Future … Short As It Is

I give SLM about two years. Within that time frame, it will undergo a few more radical updates, a whole slew of “little fixes” … and it will continually dig deeper and deeper into the sucking mud that drags all software products to their death. The LL Development Team has condemned all their work to this fate solely because they refused to honestly evaluate what was needed and why. So when they redo it again in two years, we will hope they have learned from this exercise and do it as an open development project that seeks and includes features and goals from every level of user.

But truthfully … I doubt they will. And that makes me almost as sad as it makes me angry.


Visit the DGP4SL Store on SL Marketplace

Comments

2 Comments on Second Life’s Marketplace – A Post-Mortem

  1. Kitty on Mon, 11th Oct 2010 10:26 AM
  2. You know, I see you write a lot. A *LOT*. Pretty words too, sometimes even with little elegant serif curves. But, I have to ask, what are you *doing*? I don’t see a lot that indicates to me that you’re really an expert on the subjects you write about: usability, modern marketplace design, etc. What I do see are some simple web services, basic sl goodies and a minimalist blog circa 2001.

    It’s rather obvious at this point that LL isn’t interested in your offers of help, so, why not build your own marketplace? Show them how it’s done. SL Marketplace isn’t perfect, but it does a lot right. So in my eyes you really need to put up something aside from Wall Of Text +1. Surely with all the malcontents out there, you can scrape together 15k for a G7 and a year of colo.

    In general I like you, and I have nothing against you as a person whatsoever. A lot of your claims come off as emotionally based, and spurious in the extreme, as they imply knowledge you cannot possible have. That more than anything annoys *me*, primarily due to the general reason-ability of the voice associated with your writing, and its saliency. It has a Rand-esque divergence from logic at the very last second.

  3. Darrius Gothly on Mon, 11th Oct 2010 10:50 AM
  4. @Kitty – Thank you for your opinion. I did work on a competitor site to XStreet for some time. Various personal issues took me away from it. Since then, LL has locked up the market by being the only website capable of sharing currencies between Second Life and their online sales, so any competitor is effectively locked out. It wouldn’t matter if I raised $500K or more, not being able to cross that currency bridge is a deal killer from the start. I won’t waste my time or money (or other people’s money) even trying.

    My complaints about LL and their decisions have everything to do with how I have done similar or identical things in the past. The mere fact that you think it’s not possible for one person to do what I have done doesn’t alter the reality. I have done all I’ve said, and I do post from a position of experience.

    I do indeed use emotions as tools in my posts. I learned long ago that motivation comes not just from cold logic or just from unfettered emotion, but from a mix of the two. I strive to keep them balanced, but being human I fall down sometimes too. Meh .. I learn and do better next time.

    Second Life is populated by some very VERY smart people. It’s developed by some smart folks too. But it galls me when I see the development side completely disregarding the residents and their experience/knowledge. No, they don’t have to do everything the residents tell them to do. But by the same token they need to not tune them out either. Smart people don’t like being made to look like fools, yet every time LL disregards their carefully chosen and presented suggestions, that is the net effect.

    Your comment is no different than many of the posts, comments, emails and feedback given to LL every day. Yet compare my reaction to you with LL’s reaction to the residents. See the striking difference? There ya go .. proof it is possible.