traffic and what my ticket result was

Sales to visitors ratio is important, but not as important as getting a lot of people into the store. If one person visits per day and they each buy one item, then that ratio is pretty darned good, to say the least, but it's only one sale a day, and isn't anywhere near as good as the other store that get 100 visitors a day, sells only 5 items a day, and has a very poor sales to visitors ratio.

I do know what you mean, Ciaran, but it isn't the most important metric, imo.

Tiny update about the new Traffic policy, from Jack's office hour today (transcript at https://wiki.secondlife.com/wiki/User:Qie_Niangao/Jack_Linden_Office_Hour_Transcript/2009-03-05):
So bots will be at least part of the discussion, as if there were ever any doubt of that. The bot data should be interesting.

You don't rent in malls Phil It becomes a much more important metric then, this is partially what I mean when I say all parcels aren't equal. Chasing the traffic number only can take your eye off the ball of seeking some balance.

Adding to Qie's quote that they will be blogging soon about search/traffic, unlike Qie, I always forget to take chat logs but it was mentioned at Grant Linden's office hour tonight.

Thank you for the update, Qie. I understood that the topic will be thrown open for discussion when it's posted, and I hope it is or they may not become aware of some very important thoughts about what will happen if they simply ban traffic bots.

This is interesting. The traffic test I'm doing involves a piece of land with ban lines up and me standing on it for 125 minutes. It started with a traffic figure of 1. It's almost 24 hours since I started to stand on it and it's still 1, but my store's traffic changed during that time.

The 1 could have already been counted for the 24 hour period when I put the ban lines up, so that's not of interest just now. What's interesting is that my time on the parcel hasn't been counted at all. I'm the owner of the parcel - it's not group-owned. I'm assuming that the public traffic numbers are updated grid-wide at around the same time, and I'm wondering if the owner's time on a parcel is not counted for traffic.

I'll wait the full 24 hours though before continuing the test. Sling used an alt and the number there changed to 0, but I don't know what sort of alt it was. I'll do the test again (after the 24 hours is up, in case the updates are spread over a lot of time). I'll use an alt for it, but not one that I paid for back when we had to pay for them. I'll use one that I know is counting for traffic. I'll also remove the ban lines in case traffic is no longer counted for parcels with ban lines.

The last time I tested it was a year ago - maybe more. It was simply a case of putting ban lines up, standing on the parcel for 5 minutes, and looking at the updated figure some time later. Sure enough, it was 5. It had been 0. I did it again with 2 of us standing on the parcel for a few minutes, and again the updated number was exactly as expected for minutes per av on the parcel. That parcel was group owned so it's owner wasn't the name of an av like the currest test parcel is.

Update: less than 24 hours later, my traffic has fallen back to 0.

Were you on the parcel in the period for which 0 traffic now shows? Is the parcel owned by you or group-owned? Were you on the parcel with the av that own's it or with an alt?

(Read my previous post to understand what I'm getting at.)

Well seeing most of traffic is bots ant they are the majority of residents online, thay could hardley be excluded

Yessss!


And the bots shout:

NO VEXATION WITHOUT REPRESENTATION!!!

How did your test turn out, Sling. I think your 24 hour period should have been up already.

My alt has been on the parcel for 48 hours now.
It's group owned and she is a member of the group (but not an owner of the group).
I'm using MetaBolt to run her.

She happens to be an 'official' alt - There was a US$9 once-off charge for declared alts at the time her account was created.

There are no ban lines up. It's a very quiet area and I'm not interested in anything but a 'close enough'. It's beside a road and river, but I have nothing at ground level so there's nothing to attract visitors - plus I don't want to put up banlines.

When she started, the traffic shown at the time was 3
On the next change - at which stage she had been there for about 11 hours, the traffic changed to 0.
On the latest change - she had been there for about 11 + 24 hours, the traffic changed to 1259 - 21.93 hours
I'll leave her there for another 24 hours and see if the sampling and update time windows give her a full 24 hours of traffic.


The 'about' uncertainty is due to the fact that the update for the parcel/sim happens in the wee hours of my RL morning. I happend to get an approx fix on the time of the first change as some RL partying allowed me to place it between two times.

If I were in the bot business and therefore seriously into the task of finding out what was going on, I'd use a grid of 16s with a bot in each one for different time windows. I'd also do a custom bot to get an exact fix on the update time.

yes, on and off for a few hours

owned by me

with the av that owns it

That adds to me wondering if owners are not counted for traffic.

Nah that's not true, I setup some new vendors in a store the other day and my traffic increased.

Then again my land is group owned.

Personally I just think they've fiddled with something with regard to av's who stay on one parcel all the time.

I have avs that stay on the pacel 24/7 and don't move at all, and they count every bit as much as they ever did. I meant owners by name - not group owned land.

/me giggles...it appears the traffic count is in fact some secret recipe....prob has eye of newt in it and puppy dog tails...

Travails with my alt



Traffic on the parcel is now 1436 - equivalent to 24 hours.

It seems that Traffic is simply avatar minutes.



Because of uncertainty about the timing of the measurement and the delay between the measurement and the update of the parcel info, people might get confusing results from short tests that happened to span the measurement event.


SO:


That parcel has a traffic score of 1436. The alt popped out for a few minutes, so maybe that would be 1440.
10 alts sitting up there would rank me in the middle of page 2 of a Search Places for "Furniture"
15 would put me at the bottom of page 1 in Furniture - and mid-way for Homes/Houses
20 would get me nearly up to the to top 50 in Furniture

15 gets me up in the top 50 in Rent / Rentals / Land


Cost = zero
Effort = zero
Kewlness = zero
Cleverness = zero
Talent = zero
Ethics = zero
Honesty = zero


Jack Linden is due to blog on the topic real soon now.
It would be nice if they actually did something effective.

Slng, I certainly appreciate you conducting and reporting on this experiment. That said, to truly put to rest this "complicated formula" BS we need an alt that spends some time on the test parcel and a bunch of time on other parcels presumably in the way Phil was doing by putting himself on the parcel for a measured amount of time and then going about his nefarious bot running activities the rest of the day. We might find then that the test parcel gets more or maybe less than 1 point per minute.

Either way, it's clear that for the bot usage this 1 point per avatar per minute calculation is either exactly right or close enough that it doesn't matter since traffic bots are unlikely to change parcels.

Sling. Is the parcel group-owned? The reason I ask is ...

I've just had another puzzling result for my test. The parcel is now 0 - down from 1 - and yet I used an alt that does count on my store's parcel. She was on the test parcel for 87 minutes during the 24 hour period that the 0 traffic figure represents, but she hasn't been counted at all, and I'm thinking it's because she's "related" to the parcel owner - me - not a group. I've just deeded the parcel to the group and set her up again to see what happens.

I've never heard that land owners and their alts are not counted and, if it's true, it's interesting, and it would also account for Spinster's traffic figues. It's not because the parcel isn't set to show in search because a couple of neighboring parcels show traffic and are not set to show in search (they are group-owned). Also the test parcel showed traffic of ~27 a few days ago, after I'd been there with a friend. Between the two of us, and the time we spent there, ~27 was low.

Right now, I'd put money on owners and their recognised alts not counting for traffic. I'm inclined to doubt that the owner(s) of a group that owns the land is/are counted, but that can be tested seperately.

It's a long time ago now, and something may have changed, although I don't think so. But when I first did the tests, that's exactly what happened. First I stood on the test parcel for a few minutes and then I stayed in SL but in other places, and not for 24 hours. The second test was two of us doing exactly the same thing - staying in SL afterwards but not for 24 hours. Both results produced the exact avatar minutes as traffic. That parcel was group-owned.

I'm trying to repeat it but it's looking an awful lot like owners and their alts are not counted. I'm now repeating the test after deeding the land to the group.

The parcel in my test was group-owned.
The alt was a member of the group, but not an owner.
The alt is or certainly was linked to my account (a group owner) as it was declared as such when it was created and a US$9 fee paid for the account.
I could try it with an alt created after LL removed the requirement/ability for the user to flag an alt account on signup - but I couldn't be arsed really

The parcel is not flagged for Show in Search, but I don't think that affects traffic in any way.

Exclusion of owner from traffic count would actually make sense if you accept that the point of traffic was to measure actual 'real-life' avatars *visiting* the parcel.
The only possible rationale for making traffic into a search ranking criterion would be that it was counting outsiders coming into a parcel - ie. "popularity".



It could be that an avatars could gain extra points by moving between parcels.
If the avatar gets the minute point for just being on the parcel for part of a minute, then it would be worth having the bots constantly TPing between parcels.
If 10-second visits were possible, then one could have 60 unique bots on a parcel in a single minute, but only taking up 10 avatar slots.

If the partial-minute thing actually applies, then this would be a big advantage for the traffic-for-hire / traffic injection operations.

Could this be the real reason that Jack Linden said "We especially want to look at the people who are selling traffic via piles of bots. That is clearly not cool."
If bots are being frantically TPed between parcels, the load is going to become significant as the practice grows.


I despair of LL though.
"That is clearly not cool."

How cool is running your own traffic bots? It seems that Jack thinks that the uncool is not so clear. Why? They only care about database load. All that ethics stuff is irrelevant to them?

Ty Sling. I agree that owners not counting makes sense.

From my tests a long time ago, I very much doubt that flitting between parcels as you described would work. The reason I think that is because I used a clock with a second hand to time my stays on the parcel, and to tell the other person when to move off it. I allowed full 60 second sweeps to make sure that full minutes had passed before moving off. Naturally, we got off the parcel a few seconds after the full sweeps but those extra seconds didn't count for another minute.

Someone wrote in the wiki that relogging into the sim can increase the maximum - not moving to another parcel within the sim. That may work because of how the current, unseen count is stored. I would think that each sim stores the current counts for its parcels and is polled for them by the central system once every 24 hours. The sim would then reset the current count. So, if the sim stores the current count, it may be that it's been written in such a way that simply arriving in the sim on a parcel counts 1 for that parcel and, if it does, then sim-hopping could be used to increase the daily 1440.

I agree with you about that too. But we didn't really need that statement from Jack to know what LL's view is. The sheer amount of time that traffic bots have been on the go speaks volumes in itself. I have no faith in them coming up with anything satisfactory. I have much more faith in them coming up with something that makes things worse, from both a system resources point of view and from a search results point of view. I have complete faith that they will come up with something to try and appease the anit-bots and yet keep their own numbers largely intact.

Access restricted sim and I get traffic for being around my house so unless friends are throwing parties as soon as I log off, group owners' traffic does count on group owned land.

(Edited to add that just in case some people aren't aware, the option to turn traffic on/off per sim has been there for forever now so getting rid of traffic would literally be just throwing a switch)

If you could do this at parcel level it might be useful, there's no need for residential plots to count traffic other than for checking they're not actually secretly running a club, which in all reality you should be able to find out anyway.