I'd hate to say that he faked it. But it really does seem like he did. I'll have to check the pre-launch emails, but I think he even warned that it might. The thing is, these days people have so many server options to prevent this sort of thing - they can rent cloud-computing that scales as necessary on the fly. Why would someone with the product launch experience of Jeff Walker have his own server sitting in a room where he has to go and press the power button to restart it!? It makes no sense. He claims now that it was all very innocent, and oh, because of the crash, he's offering the product again for those who missed out! The course looks great, and I would consider buying it, but I'm a bit concerned about buying a product launch formula from a guru who can't figure out how to a) handle the technical requirements for the launch himself or b) hire someone who can.

Confidence Shaken.

I don't know Condredge… I'm inclined to believe him.

I had a launch a few years ago (March 2006?) that created $570,000 in sales in under 24 hours. (Not for me unfortunately - it was for a client.)

THAT site also crashed.

Although the site was running on a pretty tripped-up dedicated box, it was ASP running on an Access database - it was never designed for that kind of heavy-duty battering.

The launch went live, the first 2 minutes we had around 350 people online and hitting the ordering page. The server crashed, phones lit up, and "Plan B" went into action…

By the time we counted the orders from "Plan B", we found that we'd sold out in the first 37 minutes… (I say "$570,000 in under 24 hours" because "$570,000 in 37 minutes" sounds too unbelievable)

That's some serious server load!

But how do you scale for a launch like that ahead of time?

Before the Launch Countdown ends, you don't know what your conversion rates will be - they could be anything from 0 to 100%. There's a big difference between the computing power you need for both - and a huge difference in the costs.

Plus, if we DID want to scale it up, we'd need to either do a complete back-end redevelopment (several months, several hundred thousand dollars)…

Or some minor redevelopments so that the site could run on multiple servers, and look for a hosting provider who would give us the "cloud-computing" on a short-term basis… Try explaining to them: "No we don't need a 12 month contract… we need a 12-MINUTE contract"

And even then, there are credit card gateway load issues, DNS issues, database integity issues and more.

It's not easy, and it's not cheap.

Uber-successful product launches are stress-testing to the extreme. They're "invited" coordinated Denial-Of-Services attacks.

Even top tech guys lose hair over these issues - they could spend months preparing for every contingency… and then the marketing flops, and there's a big financial black hole under "technology" - depending on how secure you wanted to be, the figure could easily be $150,000 or more!

The theory is a lot easier than the practice.

I learned my lesson the second time… Instead of one big launch, we did 3 VIP "Early-Bird" launches to 3 different databases before the official launches, reducing the load on the server.

We ended up making just shy of $1,000,000 the second time around…

But if you're the product launch expert… which of these scenarios would you prefer… $1,000,000 over 1 month… Or $570,000 in 37 minutes, and demand was SO HUGE the server crashed.

I know which one I'd prefer to use for my next product launch.

Sorry John - I hijacked your post here.

But just finally - I never addressed your point:

"Makes you wonder how companies like Wal Mart and Target can keep their websites up and running with all the buying traffic they must get…"

In my humble opinion, this is easier to scale than a one-off product launch because it's long-term. You're dealing with a 1,000 lb Gorilla - not a Chimpanzee with a 990 lb finger.

No problem Brent, it wasn't a hijacking as far as I'm concerned - you made some good points.

My comment about Wal Mart and Target was more tongue-in-cheek than anything but that's a good point that dealing with ongoing traffic is easier to plan for than a short-term spike.

I've seen it happen on the backend as well as the sales process. When Frank Kern's Mass Control went live, the membership area crashed the server with all the people logging in at once.

These are people that already paid a lot of money to join, so there's no value whatsoever in having the membership section crash - it's only going to annoy paying customers.

The cynic in me wants to say that it's hard to believe that the guys that have had these big launches before can't plan ahead to avoid these crashes. But your point about hosting companies not understanding the need for this kind of backbone for a matter of hours rather than months is a good one.

Even the high-profile marketers typically only plan big launches once, maybe twice a year. That's a lot of wasted server power if you have to sign a longer-term contract.