The "Our Servers Crashed" Launch/JV Technique?

[Paul Myers]

Chris,

You don't really need me to quote the very products that teach it, do you?

Yeah. Please do.

Who wrote the one you just quoted from, by the way? This sounds like a person to be avoided, unless that's the kind of parody it seems to be.

It's funny that no-one bothers trying to answer my very simple questions. So many people are ready to accuse folks of lying, but to explain why? Nah. Too much trouble.

Roger,

Yeah. I get that they're cynical. I suggest that the cynicism is more a function of clueless people making baseless accusations and not getting called on them than it is of any bad practices that may occur.


Paul

[davemiz]

paul you're dead on.... cynical people making clueless statements.

[mbealmear]

It does happen all the time with big launches like what these guys have.
You have to consider that most of there JV partners and top affiliates have thousands and thousands of people on their email lists that have been notified in advance of the product and are already ready to buy, not to mention all of their other traffic strategies.
All of those people hitting a website at one time will bring it down in a heartbeat!
Even though they are "gurus" and they should "know" what the launch will be like..... even they cannot predict how many people will try to hit their website at one time!

It does baffle me a little why the 'it's not a tactic' side (which obviously contains some of the sharpest minds) keep repeating the same mantra and appearing to fail to understand the other side's point, which is summarized here -

Quote:
Because you'll make even more if people feel they can't get it when they want it, and you'll almost automatically take anyone who was on the fence about purchasing it to wanting it even more and maybe into a committed buyer, that's it.

The reason why that statement is ignored is because the statement itself is bogus. There is no real marketer that is going to put his/her name on the line just to "try" some hyped up crap that could cost him/her their business.

We assume things with that statement above but let's look at facts..

Server Crashes and customer can't order, they get upset and go someplace else and it really is that simple. Customer is not going to listen to a bunch of hyped up crap to go back to the site after getting another email.

Few Facts

* Customer 90% of the time can go do a google search and find a very similar product. (the place they buy from has no server crashes)

* Customer feels that if you cant handle your own business then why are they going to spend money with you. (customer goes someplace else where a business can handle the traffic)

* Customer feels you dont know what you are doing so why trust you with what your product claims to be able to do. (again customer goes someplace else)

If the product being sold has anything to do with website development, scripts, servers, online tools, software (online or off), info products on how to run a business, and etc... Then the above facts will come into play with those customers.

Thus the marketer losing many many sales... Those hanging on the fence as it is stated will be very few. If the product is very unique and been hyped up a great deal then you mayget more hanging on the fence waiting for that 2nd email but the fact is the damage it will do to your business isnot worth even trying such a tactic ..

For those of you that don't understand what a real server crash is I suggest doing some research before making blatant claims that it is a tactic. Cold hard fact is a server crashing is not a fun day at the naked beach... It is a freaking nightmare in hell with the devils bitch daughter whipping the hell out of you..

Those marketers that have had real server crashes know exactly what I am talking about..

James

[Daniel Molano]

Originally Posted by digabot

So, has anyone noticed that launches these days (PPC Classroom, Continuity Blueprint, etc) all seem to be using the "we had so much interest we couldn't handle the volume" approach? The servers went down, blah, blah, blah. This is usually followed by additional "bonuses" that are simply ways for them to allow their friends to make offers to these same prospects/buyers.

Any drawbacks to using this launch technique?

Big launches are bound to crash ANY server, especially if it's shared hosting.

However, if you are running a VPS or dedicated server simply contact the service provider prior to launch and tell them that day x will receive crazy traffic and that at all costs the website must remain up, simple.

[flnz400]

Anticipation IS a tactic, for the record.

Load balancing anyone?

[Lance K]

Originally Posted by haridasz

Defiantely a case of curring corners.

Based on experience, these marketing experts and their 'gang' would know what to expect and what not to expect.

Devil's advocate here...

But if you believe they're so adept at psychology and marketing tricks, then you would surly think that they're capable of pulling them off without their prospects figuring them out.

[John_Reese]

This wasn't a marketing tactic in the case of Ryan Deiss. Ryan is a close friend and I know what the situation was. Ryan was actually bummed out because it killed momentum and kept them from taking orders when people were ready to buy. Anyone that has seen data from any of these big launches knows that a huge percentage of the launch sales occur within the first 60-120 minutes. You'd be an idiot to claim server problems at the start of the launch. Ryan lost a lot of money because of the server problem. His host made some promises that apparently weren't accurate.

Garrie - you're right that I had stated in the past that my million dollar day back in 2004 benefitted from a server issue and the need for people to remail. But my situation was different. The inline videos (that would play right on the sales letter) weren't working by the time we needed to launch (we had a Flash expert working on them all night and he still couldn't solve a small issue that we had) -- this was at a time when no one was using inline videos in the sales letter (that don't take you off the page) like is so common today. So we launched the sales letter without the videos working. After another 4 hours we got the videos working and it WAS a great reason why to get affiliates to promote it again.

But as you may or may not have noticed, the TS2 launch went off without a hiccup and that was one of the biggest launches in the history of the industry. We were using a load balanced array of 10 servers and it handled all the traffic (and database calls) without a problem.

So I will say it... any marketer that would actually try and use a "the server crashed" excuse on purpose is a total idiot that is only losing money. It is NOT a good marketing strategy that will net the highest amount of sales.

-John Reese

[othellotech]

Originally Posted by Chris Monty

Email 1: "We're open!"
Email 2: "The Server Crashed!! OMG!"
Email 3: "Ok we're back up and there are only 24 spots left!"
Email 4: "We're closed!!! Sold Out!"
Email 5: [ Explanation about overwhelming traffic and how it broke the servers ]
Email 6: [The 'Honesty Email'] "Ok just to show you how honest we are we're going to admit that 94 people have quit [insert uber-product name here] and these spots are NOW AVAILABLE!"
Email 7: "17 of Our New Servers Crashed!"
Email 8: "All spots are taken"
Email 9: "Our top member just made $46,394 in the first 4 days alone, and he's blind!!"
Email 10: "Act fast there are 15 spots available"

I've not laughed about IM so much in ages - thank you for posting the play-book

[Chris Monty]

Originally Posted by davemiz

paul you're dead on.... cynical people making clueless statements.

Actually, I'm about as optimistic as they come.

To be fair, I will say that I don't think that ALL server crashes are marketing ploys.

There are absolutely times when your server just can't handle it. I've experienced myself, albeit on a much smaller scale. It's frustrating and it's a pain in the ass.

I still think, though, that there are plenty of marketers who like to build up the hype as well as marketers who still teach it (the server crash tactic).

You can usually get a sense for when the crashes are real and when they are not just by the preliminary JV material sent out by the marketer. These days I can smell a shark a mile away.

[GarrieWilson]

But as you may or may not have noticed, the TS2 launch went off without a hiccup and that was one of the biggest launches in the history of the industry. We were using a load balanced array of 10 servers and it handled all the traffic (and database calls) without a problem.

Thank you, John.

For showing that *is* possible to do a massive launch w/out crashing.

Maybe others will follow your lead and invest the time and money for future launches.

Garrie

PS I honestly didn't know though. Trying to keep my nose to myself as much as possible.

[Chris Lockwood]

It's amazing how many people who have never been involved with a big launch are so sure that a server crash must be a lie.

If it were my launch I'd much prefer the server stay running- it makes no sense to intentionally fake a crash. The people who are accused of that are already getting lots of publicity and don't need to pull a stunt to get more.

[Cash Blasters Online]

what it crashed.......OMG !!

[wolfmonk]

A division of Rackspace called SliceHost.com offers a cost-effective solution, anyone tried them?

I asked Rackspace what if I get on the first page of Digg, and they suggested Slicehost.com.

They use the virtualization software to reserve memory and CPU slots, then employ a custom built backend to manage their slices.

256slice = shared hosting
1024slice = dedicated server

If anyone tried them, please share your experience

Thanks

[MikeAmbrosio]

Originally Posted by Killer Joe

Paul,

Thanks for that info. Not my field of expertise.

But I think even if you solved that problem you would still see all the crashes.

And here's why...

As a marketer if I "crash" my servers I have a somewhat "legitamate" reason to contact my list and give them "up to date" progress reports that allow me to put my product in front of them again.

And again, and again.

Each time I watch one of these launches unfold, the email updates come in at a frantic pace. If this pace was used under normal circumstances the unsubscribe rate would go through the ceiling.

So this marketing tactic is not only about social proof, it is also a ploy to get the list to allow the promoter a chance to bombard those interested in the product with additional reminders.

Not necessarily a bad tactic. I think the proof of its worth is in the continued usage of the ploy.

Is it overplayed?

I'm tired of it, but I'm not the one benefiting from it.

KJ

Now, no offense, but this is just silly talk.

How oh HOW would crashing your server - not allowing people to buy your product at the moment they want to pay for it, then have them leave, just to have to mail them agaiin, beg them to come back and try again - could this POSSIBLY be a benefit?

During Mike's launch, he sat there on Ustream for 6 hours total, keeping over 4,000 viewers (buyers) waiting and you think that was some sort of ploy? LOL!

Think about that long and hard. Seriously.

Is this a tactic YOU would employ?

Sometimes it amazes me that people actually believe that this is done on purpose.

And as for comments about how these guys should be better prepared, etc., well, even Mike agreed with that one. But as he also stated during all of the hoopla on his launch this week, his last 3 or 4 went off without a hitch. He (we) thoroughly underestimated the amount of people that would hit the buy button at the same moment.

What Paul Myers wrote in his post about database errors, that is exactly what happened here. And it snowballed from there. During that downtime there was a team of 4, plus Kiosk (who, btw, were very helpful AND responsive to Mike's needs. We had Joel himself on the phone) and all of them worked very hard to get things back on track.

In the end, it was frustrating to have to go through this. But we all learned some valuable lessons that will be applied next time around.

So please - enough about "staged server crashes". Trust me when I tell you - the extra emails are NOT a benefit in a case like this. Think about that logically.



Mike

[kitch29]

"If an email was sent out explaining the issue instead of being "hyped" like most other messages, it would be believable." I agree

[Tony Vercetti]

I dont think they would be this clueless as to the bandwith that their server can handle. It is definitely a marketing tactic, to make you think that so many people are getting the product, why shouldn't you. Doesn't it seem a bit strange to anyone that despite the site having crashed... the payment page seems to operate 100% ok in those cases?

[stevenh512]

Originally Posted by MikeAmbrosio

During Mike's launch, he sat there on Ustream for 6 hours total, keeping over 4,000 viewers (buyers) waiting and you think that was some sort of ploy? LOL!

You're one of the people here who knows Mike a lot better than I do (I only really know him by his products and an email exchange I had with him a few years back).. so I'll take your word for it. But you have to admit, the fact that he joked about "programming the server to crash" on Twitter several days before the launch does lend a little credibility to that idea in the minds of those who want to assume it was a marketing tactic.

I won't say whether the server crash in that particular launch was a marketing tactic or not because I don't know. You were there, I wasn't, so like I said I'll take your word for it.. but I do know this much. Servers crash, it's a fact of life, no computer or piece of software is perfect. There are also a few marketers out there who have been known to fake a server crash as some kind of marketing tactic. It's never safe to assume that a server crash on a big launch isn't the real deal, but I don't think it's always safe to assume that it is either.

[eholmlund]

Originally Posted by John_Reese

So I will say it... any marketer that would actually try and use a "the server crashed" excuse on purpose is a total idiot that is only losing money. It is NOT a good marketing strategy that will net the highest amount of sales.

Agreed.

I'm not sure if anyone has mentioned this, but another reason no sane marketer would fake a server crash is because of their affiliate/jv partners.

Not only does a server crash piss off the potential customers who are waiting; it pisses off the affiliates who are promoting the launch. And most of you understand that your affiliates are one of your biggest assets, and you do NOT want to alienate them.

When a server crash happens it makes affiliates look like fools in the eyes of THEIR lists, because they're promoting a site that is down, and they end up having to send more emails than they had planned, etc.

No one wants to throw their friends under the bus like that, which is why they would not manufacture or even hope for a server crash.

I have been part of a million dollar launch, and I know the strain it can put on servers. I watched my server load go up from 0.1 to over 100 during the minute we launched. Fortunately we were able to hold it together, but I can certainly understand how servers crash.

That said, it IS possible to manage traffic and balance the load.

In Mike's case, I wish he had been better prepared, but he underestimated the traffic.

I can't say that it's *never* been used as a tactic, and it wouldn't surprise me if someone had tried it.

But in this case, and most others, it is clearly not intentional.

[Black Hat Cat]

Not only does a server crash piss off the potential customers who are waiting; it pisses off the affiliates who are promoting the launch. And most of you understand that your affiliates are one of your biggest assets, and you do NOT want to alienate them.

And that's what is so ironic about this lame argument. People come here and bitch about crashed servers and how much it pisses them off when it happens, then in the next sentence argue that it's a ploy to make people want to buy more, despite just saying that it pisses them off and makes them less likely to buy, lol.

But you have to admit, the fact that he joked about "programming the server to crash" on Twitter several days before the launch does lend a little credibility to that idea in the minds of those who want to assume it was a marketing tactic.

Well, yeah, to those folks a few fries short of a Happy Meal. To those few people, it's also evidence that bigfoot exists. The rest of us who have more than two brain cells to rub together actually got the joke.

[jasondinner]

Originally Posted by digabot

Sorry folks, this is a straight up launch tactic. PPC Classroom had a 6 server cluster. They had around 5,000 orders with staggered (private) onsales. No way that crashes that system.

In the case of Continuity Blueprint - People Crashing a couple of servers to get in on a (thin) $2000 product. Please. This is an excuse so that everyone promoting the program has additional chances to remail their lists. Also allows them to pool bonuses so that they can grab/share prospects for their own lists.

I was sitting right next to him (Mike) the entire time.

It wasn't staged.

Thousands of people clicking the add to cart button and refreshing the order page
over and over and over again can and does cause issues like the one we just saw.

Ryan Deiss crashed his entire cart when he flipped the switch for Continuity Blueprint

Needless to say, I lost a bunch of sales as both of their affiliates.

Why would he want to purposely take money out of his own and his affiliate's pockets?

[jasondinner]

Bigfoot DOES exist!!! LOL

[stevenh512]

Originally Posted by jasondinner

It wasn't staged.

Ok, now that's 2 people who know Mike and at least one I know for sure was sitting in the same room as him the whole time (I remember seeing you on the ustream jason, don't remember if I saw Mike A. or not lol). They both say it wasn't staged. They were there (at least one of them for sure), most of the rest of us weren't. I'm inclined to believe them, regardless of Mike F.'s joke on Twitter several days ahead of time, I seriously doubt that particular crash was staged.

[Martin.Avis]

It is interesting. Almost every big launch that has happened in the last few years has been accompanied by a server crash of some kind or another.

Poor planning? Inevitable overload due to the way traffic is created? Cynical marketing ploy?

It is very easy to accuse all and sundry of the latter but for one thing that I've yet to see anyone mention...

All of these big name marketers have staff. In some cases lots of staff. An 'engineered crash' would involve several people and would certainly be well known about by a lot more in the organization.

Organizations leak. There would have been a whistle blower by now.

I hear no whistle - and I'm certain that someone, somewhere would have let the cat out of the bag if it was wriggling.

No. The reason we see a server crash on every significant launch is not because of cynical marketing hype, but because they all follow the exact same launch formula - which has at its heart a significant flaw: driving huge traffic numbers to all arrive at the same precise instant of time. It is too powerful for the technology that currently exists.

Martin

[MaskedMarketer]

Originally Posted by digabot

Sorry folks, this is a straight up launch tactic. PPC Classroom had a 6 server cluster. They had around 5,000 orders with staggered (private) onsales. No way that crashes that system.

In the case of Continuity Blueprint - People Crashing a couple of servers to get in on a (thin) $2000 product. Please. This is an excuse so that everyone promoting the program has additional chances to remail their lists. Also allows them to pool bonuses so that they can grab/share prospects for their own lists.

I wouldn't know if its a "tactic" everytime, but didn't it also happen to eban, frank k, the AC dudes, and now mike Filsame? (and probably more)

And now Mike Filsame is extending his order buyers to more than 5k? hmmm

Obviously we would never "know" if it was setup that way, but its a tactic marketers teach.

And If your someone that's not familiar with the tactic and the people teaching it- thats not my problem.

[Martin Luxton]

Not having been responsible for a million dollar launch I just want to ask if it is a good business model.

Spending months and months and hundreds of thousands of dollars on the project just to be at the mercy of a computer crash (which the technically savvy here seem to agree can happen all too easily) sounds like madness.

I don't have figures but it would seem the money from the launch comes from establishing yourself as a guru and the massive list of buyers you build, and not from the product itself.

If that is the case does it matter if the server crashes or not?


Martin

[JayXtreme]

I really don't know about this.. I am always in two minds when this discussion comes up.

This probably stems from the fact that we can all feel that sometimes it is genuine and other times it isn't..

But I will say this.. I have a couple of decent size lists in this niche, and I wouldn't touch them with a launch (as an affiliate) right now.. it's too painful to be involved, I hate being involved.. you lose sales, subscribers and cred with your list...

I have felt this first hand...

Just Sayin'

Peace

Jay

[dmag1]

Originally Posted by wolfmonk

A division of Rackspace called SliceHost.com offers a cost-effective solution, anyone tried them?
Thanks

I haven't but one of the top bloggers in the world is a friend and partner, and he uses them. He's been on the front page of DIGG numerous times and the system has never even burped. He recently got 100,000 unique visitors over the space of a few hours and there was no problem.

The downside of Slicehost is that you need some technical knowledge to use the system. It's not like shared hosting with a simple GUI. That said, they're getting friendlier and now give links to individual cPanel licenses for folks who need them.

He swears by them.

[James Schramko]

It pisses a lot of people off. The affiliates, the merchants and the poor consumer who cops a third, fourth and fifth round of "the server is back up" emails blended with apologies...

It would be stupid to manufacture a server crash.

[wolfmonk]

Originally Posted by jacktackett

Hum, maybe an enterprising warrior could build out a system and then lease that system to the gurus for a launch - it would require some great scheduling tho - especially with the propagation delays inherent in DNS.

--Jack

Jack, how much of importance is redundant DNS in place? Is there a probability a hosting company won't be able to handle too many DNS requests?

These guys say this can be an issue: dnsmadeeasy.com

Thanks

[ExRat]

Hi,

I generally don't bother discussing stuff like this anymore, but as I was involved in this resurrected dinosaur thread previously, I might as well jump in with an observation -

Almost everyone (on the one side of opinion) is saying this (below) as if it's utterly conclusive and ends all debate -

It pisses a lot of people off.[snip]

It would be stupid to manufacture a server crash.

Let's use an analogy to make the point - email marketing (another subject that creates similar threads).

What's a marketer's priority in email marketing? Answer = sales

What's one of the by-products of this process that is accepted and dealt with as one of the consequences of achieving that main aim - pissing people off.

So the question remains - with all of these launches, do they piss so many people off with the server crash that the launch is a failure?

Based on what we have seen, the answer can only be 'no'. Just look at the other threads where everyone is discussing Mike Filsaime's recent launch - and regardless of what they are saying, one thing seems to run through the majority of posts - they all bought it.

So what on earth is so different between typical IM email marketing and typical IM product launches?

Why with launches is everyone suggesting that it would be 'stupid' to manufacture a server crash and thus, 'piss people off' yet it's standard procedure to piss people off with IM email marketing, because making sales is the main aim - and if you focus too much on not pissing people off, you will fail to make enough sales (or so it would seem, judging by 'typical IM 'churn & burn' email marketing tactics.')

Am I the only person that sees it this way?

I can think of a ton of reasons why having a server crash is a particularly good way to boost sales (at the expense of pissing off a few potential customers.)

Someone said above about Mike Filsaime 'having to do a webinar for four hours' - as if the implication was that this would be a real pain in the ass, in order to secure thousands of orders, thousands of confirmed email addresses of buying IMers, thousands of physical addresses/phone numbers? (I don't know because I didn't buy it), thousands of people with a continuity offer/upsells/OTOs put in front of them WHILE they have credit card in hand....

I think even Mike would tolerate four hours of putting himself in front of those people, communicating with them, having them hanging on, ready to buy for four hours....

If they bother to hang around or keep checking back for four hours then they are damn sure to follow through and buy the product - and add to that the fact they are probably reading/communicating with him - IE - he's breaking down barriers with them and becoming familiar with them, getting them to empathise with him due to his 'sod's law' based server issues, offering extra bonuses to make up for the wait and endearing them to him EVEN more.

I have no gripe about it as it doesn't affect me either way. If anything, he should be applauded (by other marketers) for skillfully exercising his jedi-mind-control-server-failure-launch tactics on people.

But I just don't get this 'it would be stupid to do this purposely and end up pissing people off' point of view when this is EXACTLY what they do in their email marketing.

Are they stupid for conducting their email marketing in the same fashion?

They're either stupid for one or stupid for the other in that case. Either that, or it's obviously a deliberate move, because it's so powerful and proven to achieve the main aim - making sales - regardless of a few pissed off punters.

Anyone notice how ALL of the discussion is about the server crash? How potential buyers are 'frustrated and desperate to place their order before the limited offer runs out'? What would they be discussing if there was no server crash? How great it all is? Would that type of thread not be deleted instantly here?

And no-one would dare to suggest that creating all of that desperation for a limited availability offer, which then turns into an UNlimited availablity offer BECAUSE of the server crash is in fact an absolutely brilliant and almost justifiable way of using scarcity, when there never really was any scarcity in the first place - would they?

C'mon guys. It might not be in the 'product launch formula' - but you're reading the wrong one. You need the 'guru version' not the 'ordinary punter' version.

The 'ordinary' version tells you what social proof and scarcity are - and probably offers some candy coated 1970s examples.

The 'guru' version explains in detail precisely what you are seeing here. Don't be so afraid to say it. It makes perfect sense.

Perhaps I should make the product myself and sell it to all you gurus who 'don't get it', even though it's right in front of your nose It works. The proof is in the numbers. See my sig file for a link to my 'almost reliable 99% uptime server' company - SodsLawHosting.

[jacktackett]

Originally Posted by Martin.Avis

- which has at its heart a significant flaw: driving huge traffic numbers to all arrive at the same precise instant of time. It is too powerful for the technology that currently exists.

Martin

For the most part I agree, except your last statement - its just untrue. I've seen plenty of load balanced systems in place with clustered databases, app servers, and web servers. One local company keeps its server farm in a data center I use to manage, over 12 racks crammed full of gear. They never have a crash - why? Because they have the technology AND architecture in place to handle things. And when they do a big launch - guess what - they'll rent additional equipment to handle the expected surge.

Now they do pay 100K/month to support this - but to just say there's no technology to handle this type of situation is as bad as saying a crashed server is just a marketing ploy.

respectfully,
--Jack

[wolfmonk]

Originally Posted by dmag1

I haven't but one of the top bloggers in the world is a friend and partner, and he uses them. He's been on the front page of DIGG numerous times and the system has never even burped. He recently got 100,000 unique visitors over the space of a few hours and there was no problem.

The downside of Slicehost is that you need some technical knowledge to use the system. It's not like shared hosting with a simple GUI. That said, they're getting friendlier and now give links to individual cPanel licenses for folks who need them.

He swears by them.

Thank you dmag1,

someone's first hand experience is invaluable, as we all can see in this wonderful thread.

[jacktackett]

Originally Posted by wolfmonk

Jack, how much of importance is redundant DNS in place? Is there a probability a hosting company won't be able to handle too many DNS requests?

These guys say this can be an issue: dnsmadeeasy.com

Thanks

Wolfmonk - there are several items to be careful of with DNS when managing load balanced systems or new servers (especially as I envisioned servers coming and going), among them are how long it takes for a network's name servers to recognize new information. Some providers are really quick, others (dare I mention Time Warner?) can take literally days to update their name servers. What happens is some folks can see your site - and others can't.

Internally, a load balancer (such as a BigIP or LocalDirector) will present one site to the world, but send traffic to various servers behind it based on a variety of paramters (some just do a round robin, others can analyze which systems are free and route traffic to it etc)
and some can determine if you need to maintain session information across responses. If that internal information is not right, things get real bad, real fast.

Session traffic is due to the HTTP protocol being basically stateless - one request doesn't know/remember other requests. Cookies are used to track the sessions - you can use databases, but that can slow things down. So if you engineer and app to use session information, you can't have the first bit of traffic going to one server, and the next bit going to a different server. You either have to share session information or force all traffic for one session to the same server. Which all load balancing devices can do.

Not sure if this answers your question so please feel free to ask others.

peace,
--Jack

[Martin Luxton]

Interesting followup email I have just received.

Because Mike Filsaime and his company
had problems during their launch they
decided to honor any request for their
special course deal and not limit to only
5,000 people.

Nearly *10,000* people have now taken
advantage of this amazing deal.

So how should the first 5,000 people who bought this thinking it
was time and number limited be compensated?

Free shipping?

Free first month of the magazine?

Martin

[Johnny Slater]

I think one thing that most people do not understand is that accessing a database and doing the processing required to complete an order is 100 times harder on a server than just viewing a web page.

You can have 10's of thousands of people all viewing a single web page at the same time and the server won't skip a beat but database processes are extremely limited and if you have thousands of people trying to access a database at one time it will cripple a server.

Web servers have a specific number of incoming database requests that are allowed. For shared hosting its usually around 256 or so. I've had sites shut down for having too many requests just from doing a WSO when using shared hosting so a huge launch can and sometimed does crash database servers.

As has been stated above, Mike had 4000 people on Ustream the second the site went live. Give it the benefit of the doubt and say his server could handle 1000 incoming database requests at any single time. With just the people sitting on Ustream waiting around for the chance to order it doesn't take a huge leap to figure out that you are getting more processing requests at once than the server can handle.

Now, if all those orders come in at once it will cripple the database and crash the server. Even if the server doesn't crash, the sales page is database driven so it is now down. Orders can't be taken because the database can't keep up with all the requests it's being hit with.

Server crashes are a reality and are very easy to have happen.

[James Schramko]

They are selling 15,000

[James Schramko]

Originally Posted by ExRat

Hi,

What's a marketer's priority in email marketing? Answer = sales

I hear you Roger, I said both those things (it would be stupid and it pisses people off) but I am not saying they did not do it on purpose. I only said it would be stupid. I still think it pisses people off.

My policy is to build relationships for the long term.

Pissing people off is never a good long term strategy.

If the day ever comes where the influx of new-to-web people slackens off a little the market will change and there will be some marketers who will not be popular.

[wolfmonk]

Originally Posted by jacktackett

Wolfmonk - there are several items to be careful of with DNS when managing load balanced systems or new servers (especially as I envisioned servers coming and going), among them are how long it takes for a network's name servers to recognize new information.

[...]

Not sure if this answers your question so please feel free to ask others.

peace,
--Jack

Thanks Jack, this clears things up a little. What I'd really like to know is shall I order a DNS services from a company like dnsmadeeasy.com with a clustered solution from slicehost.com so that I can reboot a slice without my DNS affected.

Actually, slicehost recommends dnsmadeeasy.com, I just don't grasp it why exactly I need a third-party DNS provider.

Thanks!

[Adaptive]

I can solve this problem, but funding the solution is a chicken and egg conundrum.

My professional experience includes many, many years programming enormous in-house data systems in retail and banking. By enormous, I mean thousands of simultaneous users hitting hundreds of servers, nonstop, all day every day, from multiple time zones.

ALL of the technical problems have been solved, many times. But the answers are locked behind corporate firewalls of the largest businesses, who each made their own version. And those answers are viewed as trade secrets.

Very large companies absolutely must have these kinds of systems to keep the business running. A colleague and I figured out that the cost of a server doesn't matter any more once it processes a million dollars a day of transactions.

Because there is no off the shelf solution for scalability, these large companies all build essentially the same system in house. Or they pay major consulting companies to build yet another version of the same system, and wind up hiring an in house staff anyway to do ongoing customizations.

Setting up such a system is tremendously complicated and expensive.

First, there's the need to have an abundant supply of servers. As others have noted, cloud based computing services such as those from Amazon make it possible to rent a server on demand. This moves the fixed cost into a variable cost.

However, there is still the need to build the software that links the servers together and keeps them communicating. There are several solutions for load balancing web servers, and several solutions for scaling database clusters. But there is no complete, "soup to nuts" middleware that automatically balances everything, end to end. Nobody has anything close to "click here to upload your web site AND DATABASE into our cloud, and we'll scale it for you."

And there are no easy answers, otherwise Amazon would already offer that particular "click here" button for your product launch.

Creating this software in the middle of all the servers - middleware - takes ten man-years. If the company started small, one guy working and working at it will eventually get it done in ten years. Or a team of ten very sharp, proficient, well cooperative people can get it done in about a year.

These would be people with 10+ years experience, who would need to be paid around six figures in salary, plus benefits. So it's two million dollars a year for the development team that makes the business possible.

Adding another hundred programmers doesn't mean you get the system in a day. Beyond about a dozen people, you need to split into multiple teams and have multiple managers and project managers coordinating everything. This means a potential for bureaucrats to get entrenched, which is why I turned to Internet Marketing to learn how to leave the corporate world behind me.

Once the middleware is set up, it still needs to be customized for each individual application.

Here is the way to the solution:
1. Get a few million dollars of funding. This is the part I'm stuck on.
2. Hire the core team: a couple each of database programmers, database administrators, web programmers, middleware programmers, testers and graphics designers. I know how to do this, if the money's available.
3. Take the first 90 days to write the spec, based on interviewing the most likely initial customers. Meanwhile the company is set up with an office, legal incorporation, letter head etc. and the geeks wire up the development servers.
4. Take the next six months to create a few versions of the system, adding more and more functionality and fixing bugs.
5. Quietly do some beta tests. Ideally this involves the kind of mega-traffic activity that infamously crashes servers. Since the client's servers are crashing anyway, route half the traffic to the new system and see if it can do better. Marketers should appreciate a split test, as should coders.
6. About a year in, release version 1.0. License it for six figures a year to large corporations that are struggling to do this work in-house, and failing because of their beauracracy. For another six figures, include months-long consulting to help them with the transition. This will let the clients cut some of their IT costs in half, although most will want to run their business on their own servers rather than trusting someone else's cloud.
7. During the next year, offer an affordable service that has a few thousand dollars initial and monthly costs, then activity-based pricing for further scalability.
8. However, if the client's code is not designed for scalability, they will need consulting help to refactor their code to properly run in a parallel, distributed environment.

Unfortunately, I don't know of a way to make the necessary technology in a single man-month or even a year of solitary hacking. If I did, I'd have already written it and offered it to all the server-crashing launchers. And unfortunately, any one piece is useless without the whole thing. If you automatically can add additional web servers but the database still crashes, what have you gained?

With the servers rented as needed from cloud companies, there is no longer a bottleneck to licensing the technology to multiple companies.

Open source is unlikely to solve this problem. The end users are businesses with a tremendous volume of business, and they're all re-inventing the same wheels as supposed proprietary advantages. The workload is different than the kinds of science and engineering applications that already have scaling solutions. It will take the creation of a new company dedicated to making a commercial solution.

I'd appreciate serious discussion of this business opportunity and challenge, either here or by private message.

Regards,
Allen

P.S. To repeat, the technical solutions already exist, the challenge is packaging them in a form that can be resold over and over to multiple clients. This is one very significant product development project.

[koolwarrior]

I say the next time someone's server crashes during a launch...don't bother buying the sh**.

If they don't even have their head on straight to have their servers READY for a big ticket launch, and they have all this friggin' money (supposedly)...then I wouldn't trust jack that they have to sell.

If they can't keep their site in order, what could that say about their content? Would it be half ass like their servers? Hmmmmmm....


I know it's a tactic. But common sense says, if their own servers and websites SUCK, especially when it's all planned out ahead of time...then their products probably follow suit, and SUCK WAY MORE than their sites, do.

[MikeAmbrosio]

Originally Posted by stevenh512

You're one of the people here who knows Mike a lot better than I do (I only really know him by his products and an email exchange I had with him a few years back).. so I'll take your word for it. But you have to admit, the fact that he joked about &quot;programming the server to crash&quot; on Twitter several days before the launch does lend a little credibility to that idea in the minds of those who want to assume it was a marketing tactic.</p>

I'm gonna let you in on a little secret...

I was there when Mike did that Twitter post. It was done on purpose in response to the cynicysm of the crashed server launch that was going on at the time. When Mike posted that and saw all theresponses, we laughed our asses off.

THEN Mike went and bought the domain. We had a whole plan in the works to do a spoof launch and crashed server, but then we simply got too busy for that.

Originally Posted by stevenh512

Ok, now that's 2 people who know Mike and at least one I know for sure was sitting in the same room as him the whole time (I remember seeing you on the ustream jason, don't remember if I saw Mike A. or not lol).

Jason was sitting to Mike's right. I was sitting to Mikes left. I had the green shirt on...

Originally Posted by ExRat

Hi,

I generally don't bother discussing stuff like this anymore, but as I was involved in this resurrected dinosaur thread previously, I might as well jump in with an observation

Almost everyone (on the one side of opinion) is saying this (below) as if it's utterly conclusive and ends all debate

Let's use an analogy to make the point - email marketing (another subject that creates similar threads).

What's a marketer's priority in email marketing? Answer = sales

What's one of the by-products of this process that is accepted and dealt with as one of the consequences of achieving that main aim - pissing people off.

So the question remains - with all of these launches, do they piss so many people off with the server crash that the launch is a failure?

Based on what we have seen, the answer can only be 'no'. Just look at the other threads where everyone is discussing Mike Filsaime's recent launch - and regardless of what they are saying, one thing seems to run through the majority of posts - they all bought it.

So <i>what on earth </i>is so different between typical IM email marketing and typical IM product launches?

Why with launches is everyone suggesting that it would be 'stupid' to manufacture a server crash and thus, 'piss people off' yet it's standard procedure to piss people off with IM email marketing, because making sales is the main aim - and if you focus too much on not pissing people off, you will fail to make enough sales (or so it would seem, judging by 'typical IM 'churn &amp; burn' email marketing tactics.')

Am I the only person that sees it this way?

I can think of a ton of reasons why having a server crash is a particularly good way to boost sales (at the expense of pissing off a few potential customers.)

Someone said above about Mike Filsaime 'having to do a webinar for four hours' - as if the implication was that this would be a real pain in the ass, in order to secure thousands of orders, thousands of confirmed email addresses of buying IMers, thousands of physical addresses/phone numbers? (I don't know because I didn't buy it), thousands of people with a continuity offer/upsells/OTOs put in front of them WHILE they have credit card in hand....

I think even Mike would tolerate four hours of putting himself in front of those people, communicating with them, having them hanging on, ready to buy for four hours....

If they bother to hang around or keep checking back for four hours then they are <i>damn sure</i> to follow through and buy the product - and add to that the fact they are probably reading/communicating with him - IE - he's breaking down barriers with them and becoming familiar with them, getting them to empathise with him due to his 'sod's law' based server issues, offering extra bonuses to make up for the wait and endearing them to him EVEN more.

<b>I have no gripe about it</b> as it doesn't affect me either way. If anything, he should be applauded (by other marketers) for skillfully exercising his jedi-mind-control-server-failure-launch tactics on people.

But I just don't get this 'it would be stupid to do this purposely and end up pissing people off' point of view when this is EXACTLY what they do in their email marketing.

Are they stupid for conducting their email marketing in the same fashion?

They're either stupid for one or stupid for the other in that case. Either that, or it's obviously a deliberate move, because it's <i>so</i> powerful and <i>proven</i> to achieve the main aim - making sales - regardless of a few pissed off punters. <img src="images/smilies/wink.gif" border="0" alt="" title="Wink" smilieid="4" class="inlineimg" />

Roger, not to split hairs or anything, but I didn't say Mike had to do a four hour webinar, etc. What I said was he sat there on his live Ustream launch broadcast for the entire time the server was down keeping the viewers informed on what was going on.

Also, let me be clear about one thing as well...

The question I am addressing is whether this was on purpose or not. To which, it was most definitely not.

The question of whether or not it's beneficial, well, I have never studied or compared launches of this calibre to determine if this way was better in the long run or not.

From a technical standpoint, from someone who was trying to get that server back on line, I can tell you the stress is definitely NOT worth it. Trust me there

At the end of the day Mike did well. But consider what he was selling too. He was not selling a $2,000 product for $2,000. He was selling it for 30 bucks. His server could have crashed for a week and he still would have done well. The pre-launch was spectacular. And it was a great value.

And Johnny Slater hit the nail on the head also. The issue was DB connections. If I recall, the server was set to 2,000 or so. We had over 4,000 hit it at pretty much the same time. We were ready for traffic - but none of us considered the database issue.

You can guarantee that will be considered for the next launch. All of these "theories" aside, Mike does NOT like the server crashing on launch day. The simpler solution is a smooth launch, sell out and that's it. The way things went here, we now have the the task of going through all of the fallout of people not getting their access info, not sure if the orders went through, etc. We love this part (not).

So, once again, consider the advantage here. From where we sit, there is NONE. It's MUCH easier to sell out in 4 hours and go home

Mike

[Martin.Avis]

Originally Posted by jacktackett

For the most part I agree, except your last statement - its just untrue. I've seen plenty of load balanced systems in place with clustered databases, app servers, and web servers. One local company keeps its server farm in a data center I use to manage, over 12 racks crammed full of gear. They never have a crash - why? Because they have the technology AND architecture in place to handle things. And when they do a big launch - guess what - they'll rent additional equipment to handle the expected surge.

Now they do pay 100K/month to support this - but to just say there's no technology to handle this type of situation is as bad as saying a crashed server is just a marketing ploy.

respectfully,
--Jack

I stand corrected jack - oh the incompleteness of communicating in text!

What I should have said was that the launch techniques most gurus are currently using are too powerful for the technology that they are currently using.

Martin

[Bill Farnham]

Originally Posted by MikeAmbrosio

Now, no offense, but this is just silly talk.

Mike

Hi Mike,

I'm glad they don't resurrect all the crap I wrote back in the '50's and the '60's. I probably would look like a bigger dummy than I really am.

I'm not going to read through all the posts here again just to condense the good points that were made by people in the know, but suffice it to say that there were enough logical points put forth that made me reconsider my viewpoint on this subject.

So the next time I get emailed over and over about some marketer having their server crash I'm just going to remember all the reasons why any big-time marketer would not want to allow his server to be overtaxed to the point of meltdown and simply conclude they lack foresight.

I'm also going to delete that person from my list.

Unless it's you...

KJ

[jacktackett]

Originally Posted by Martin.Avis

I stand corrected jack - oh the incompleteness of communicating in text!

What I should have said was that the launch techniques most gurus are currently using are too powerful for the technology that they are currently using.

Martin

In that case - I totally agree with you. You have to spend some $$ to get the right infrastructure. And as others have said here - its not just the front line web servers - you need to have the right database architecture in place. And those aren't cheap either.

best,
--Jack

[stevenh512]

Originally Posted by MikeAmbrosio

I was there when Mike did that Twitter post. It was done on purpose in response to the cynicysm of the crashed server launch that was going on at the time. When Mike posted that and saw all theresponses, we laughed our asses off.

lol.. Yeah, I assumed it was a joke when I saw it, but like I said you have to admit it does kind of lend a little credibility to the idea in the minds of those who were inclined to believe it was staged anyway.

Originally Posted by MikeAmbrosio

Jason was sitting to Mike's right. I was sitting to Mikes left. I had the green shirt on...

Ok, I remember seeing you now.. so, that's 2 people who were in the same room with Mike the whole time.

Originally Posted by MikeAmbrosio

The question of whether or not it's beneficial, well, I have never studied or compared launches of this calibre to determine if this way was better in the long run or not.

I can't say if it's actually beneficial or not, I don't know if there's been any testing to prove it one way or another.. but I know there have been a few "gurus" who have taught or used it as a marketing technique. Not saying Mike F. is one of them. In fact, as long as the server was down and with two eyewitnesses who were in the same room as him when it happened (even if you do both happen to be good friends of his.. lol) I don't believe he faked it. Most of the crashes I've seen that I suspected of being faked (or knew were faked) only lasted a few minutes to an hour, they weren't an all day event like this one.. lol

[Paul Myers]

I know it's a tactic. But common sense says, if their own servers and websites SUCK, especially when it's all planned out ahead of time...then their products probably follow suit, and SUCK WAY MORE than their sites, do.

I would like to offer you some advice. Take it or not, as you please...

The proper sequence is: Read first. Then reply.

You don't "know" it's a tactic. You believe it is, but that's an entirely different thing. No opinion can change the reality of a thing that's already occurred.

As far as correlating the product quality with technical experience... That's not exactly logical as a sweeping statement, is it? For example, if I were to release a product about the communications aspect of selling, and the server crashed, how would that suggest that the product was somehow deficient?

One of the challenges here is that very few people with the personality traits involved in being an effective marketer have any serious understanding of the technical side of things. You're talking about two different mindsets, and they rarely co-exist in the same person. The time demands involved in the two areas preclude a significant overlap of those skill sets at any high level of proficiency, for most people.

I don't know Mike Filsaime, and I didn't follow this launch at all. However, if Mike Ambrosio says it wasn't a planned tactic, I believe it wasn't a planned tactic. Full stop.

I do know a bit about some of the other launches that have encountered server problems, and I can tell you the crashes were not planned. Or fun.

I keep hearing about how someone supposedly teaches folks to do this. I'm still waiting for one of the people suggesting this to point me to the person who allegedly encourages that sort of stupidity.

I suspect I'll be waiting a long time.

And yes, it would be stupid. Getting your launch partners involved in a deliberate lie could be a crippling thing for a business, if it were to be found out.

Not worth the risk.


Paul

[MikeAmbrosio]

Originally Posted by Paul Myers

I don't know Mike Filsaime, and I didn't follow this launch at all. However, if Mike Ambrosio says it wasn't a planned tactic, I believe it wasn't a planned tactic. Full stop.

That means a lot to me, thanks

Originally Posted by Paul Myers

I keep hearing about how someone supposedly teaches folks to do this. I'm still waiting for one of the people suggesting this to point me to the person who allegedly encourages that sort of stupidity.

What I think this is is Jeff Walkers PLFormula - he apparently teaches that IF there is some sort of problem on launch day, or with a link in your email, etc. then take the opportunity to rebroadcast.

He does NOT advocate CREATING false issues. Just that you should follow up with your customers if you DO encounter issues.

Someone correct me if I am wrong here.

Mike

[Paul Myers]

Mike,

Someone correct me if I am wrong here.

You are not wrong there.

Jeff gets a kick out of the people who think this is some kind of marketing ploy.


Paul

[o_brient]

Originally Posted by Paul Myers

Mike,You are not wrong there.

Jeff gets a kick out of the people who think this is some kind of marketing ploy.


Paul

Right one and all.

My credentials - I've been in IT 20 years now and have been the following:
An Analyst/Programmer
A Software Engineer
A Senior Software Engineer
A Technical Architect and now my day job is
A Trainer in Business Process Modelling leveraging my IT skills (Love a paid captive audience - the only possible server failover is if my knees give out from an over-indulgence the night before)

My "night/weekend" job is unfolding...

Let's put the record straight on this one.

To my knowledge, Jeff Walker, Frank Kern, Mike Filsaime, John Reese, Rich Schefren and a whole who's who of Internet marketing - including the living legend that is Paul Myers DON'T know about the complexities of a fully
Reliable
Available
Scalable
Performant
Usable
Extensible
Redundant
Clustered
(And I could go on and on...)
computing infrastructure.

They rely on others for the "techie" details and each has their own computing infrastructure delivering their own businesses.

This is no problem.

As Will Rogers said (one of my favourite quotes) "We're all ignorant - only on different subjects".

They don't have to know about how to setup a LAMP system, how to put their computing needs into the Amazon Cloud using EC2/S3 or having a full on in your face clustered environment using multiple firewalls, having a DMZ or creating multiple layers to separate presentation, business or database logic.

If the aforementioned makes no sense to you then GOOD, it doesn't to the guru's either.

Why should it?

At the end of the day there are different shades of people around the world. I'm not being racist, I'm being a "scrupalist" [sic].

Some people adopt "black hat" strategies.
Some people adopt "grey hat" strategies.
Some people adopt "white hat" strategies.

White, grey/gray (case in point), black are colours/colors (case in point) which look different depending upon perspective.

Servers will go down, that is a fact.

Some people will engineer server failures to be able to report their "tsunami moment" as Jeff Walker puts it (he does not advocate doing this btw as if you do it will come back to bite you on the ASSets you thought you had) - that is a fact.

Some people will be unfortunate even if they think they have setup a system which should in theory cater for a massive influx of demand (for example, and I kid you not this is based on real world experience - a major company kept experiencing repeated outages every Friday evening after hours, long story short - a cleaner needed a hoovering point and their servers did not have an uninterruptible power supply - you guessed it, the cleaner unplugged the nearest computer/server to do their job).

You cannot legislate for everything and you can be damn sure no matter what you do there will be some detractor or other flinging guano in from the sidelines.

In fact, let's face it, many's the time such guano slinging is intentionally provoked to ellicit a response (not saying this is the case...)

Rant off.

My first name is Thomas/Tom so I sure know how to doubt but given my experience I also know you have a choice.

Believe the marketer in question or don't.

Move along regardless.