The Curse of Success


#1

Reading many of the recent conversations here has led me to reflect on what it means to succeed out of proportion to the expectations. Glowforge placed a funding goal of $100,000 on their pre-order campaign, which would have been met with 50 Basic machines. Instead, they reached over $27.9M with orders for several thousand machines. The pre-orders have grown to about 10,000 total since then. That’s on the order of 200 times the funding goal.

Tesla unveiled a prototype for the Model X in 2012, four months before starting deliveries of the the Model S. They expected to start deliveries of the X in 2013 and thought the X would be mostly a different body on top of a dual-motor Model S platform using perhaps 90% of the same parts.

Then the Model S started selling much faster than they expected. They had barely the battery supply and the factory production capacity to meet the Model S demand (with significant order-to-delivery delays), much less add a completely new model to the mix. Between Model S selling so well and the significant feature creep on the Model X, the X would not start production until September 2015. That required building a new factory line and committing to long-term large battery supply contracts.

In 2014, Tesla announced their plan to build a battery Gigafactory that, by 2020, would produce annually as many Lithium Ion batteries as the entire world produced in 2013. They also announced that they expected to be making 500,000 cars per year by 2020.

On March 31, 2016 Tesla started taking $1000 order reservations for the Tesla Model 3, which they would unveil that night. There were lines of people, many were several hundred long, at Tesla showrooms worldwide that morning. By the time the prototype cars were shown, there were over 115,000 reservations placed worldwide for this car which would start deliveries in “late 2017”. By the end of the week, Tesla announce over 373,000 reservations had been made.

This was significantly more interest than Tesla had anticipated. It led Elon Musk to famously tweet, “Definitely going to need to rethink production planning…” on April 1. A few weeks later, Tesla announced they would meet their 2020 production goals in 2018 instead. Most in the auto industry thought they were crazy.

Tesla has started production of the Model 3 ahead of schedule (something new for them), and expects to be making 5000 per week by the end of this year, and 10000 per week sometime in 2018.

Bringing this back to Glowforge…

If Glowforge had only succeeded in meeting their goal, they could probably have delivered 50 Basic units by December. They could have 3D printed the cases rather than tooling up for injection molding. They may not have encountered many of the problems that occur in only 1% of units. They would not have needed to be sure their online services could scale to the needs of 10,000 users.

The moment they blew by their pre-order goal on the first day, they knew they would need to “rethink production planning.” By the end of the campaign they were likely trying to plan out how to deliver thousands of machines rather than tens or hundreds.

Over the two years since the pre-order campaign, they have uncovered many of the one-in-a-hundred or one-in-a-thousand problems. They have discovered many things they didn’t know that they didn’t know at the outset.

In short, they’ve met the Curse of Success. Things that you can work around when hand-building small numbers of devices become major problems when trying to make thousands of those products.

But now they are shipping. There are likely hundreds of Glowforges in peoples’ homes and businesses. These machines are significantly better engineered than they would have been if only 50 had pre-sold.

None of us has enjoyed the wait, especially not Glowforge, I’m sure. But it is not at all surprising that the delay happened–especially when coupled with the Curse of Success.


#2

True. I make better pizza and bbq than anything I’ve had anywhere else (I take the best from every good one I eat and incorporate it into mine) but only because they are all one off products of my oven or smoker. I could never do that at scale. So I’ll never own a restaurant or a food truck.

Almost everything is easier when you have the freedom to concentrate on producing just one. The trick is producing that one’s features and quality a thousand times.


#3

If they had only sold 50 machines over the course of 30-45 days, the company would have immediately shut down. That number just can’t sustain a company the size of Glowforge (even the size that it was.) They certainly knew during the preorder campaign that they were at least a year away from having a shipping product. I’m sure they had an internal target that was much, much higher. They set a very low number so they could say they exceeded their “crowdfunding” goal by 20x.


#4

I would tend to agree with you. However, that still would have been maybe 1000 machines. They exceeded that by an order of magnitude. Manufacturing scale matters, and when the number required jumps by an order of magnitude, the decisions one makes on how to make things changes.

I’m sure they would still have run into things like the Power Supply issue, and other such problems. My point, though, stands that when they put together the pre-order campaign they very well may not have known how large the issues would be.

And that brings me back to a point I’ve made in other places: It seems to me that Glowforge has forecast schedules to the best of their ability with the information they had at the time.


#5

Some expectations out of @Tony’s mouth back in the day:

“Internally, we thought we could hit $2M in sales and talked about calling it a “high five success” if we hit $5M.”


#6

We’ll never know what their “real” goal was. There wouldn’t have been a company if they really expected to only sell 50 units. They knew that, of course. So I don’t think there was any re-thinking for that reason. Plus, it made it all the easier to say “Oh, look what we did! We’re so awesome that we far surpassed our goal!” Now maybe they didn’t set their sites so high. But there was a number. A known number. A number of units that’d have to be pre-ordered in order to proceed as a company.

Edit: Oh… And @pdobrien has basically already said what I just said. Ah, well… :blush:


#7

My expectation is that this wave is also hitting their tech support hard. If you are a car dealership, you might have surges in demand, but you’re basically delivering units and supporting them as you go along. Now imagine trying to deliver two years worth of pre-orders for a brand new, never been made before car over a few months. The wave of new cars out in the field, discovering problems with them, and supporting the increasing numbers of new owners would cause a significant challenge in support. There might not be enough garage space for fixing everything in a timely fashion or answering the phones in a timely manner, etc.

Bottom line: I have to give the Glowforge elves credit for tackling this challenge as well as they have, but also think we should be respectful that they have a lot on their plate in terms of support/fab/problem solving as units go into the wild. Design/production was a huge challenge. Support is going to be another huge challenge.


#8

No argument there… I’ve seen it time and again with Kickstarters I’ve backed. People who thought they would sell 100 units were planning on assembling them by hand, then they sell 1000 or more and realize there aren’t enough hours in the day.

Random problems aside, the preorder campaign ended in October and they said they were starting to ship in December. Right now, 2 full years later, they still haven’t managed to figure out how to get plastic casings without flaws. There’s no way they could have shipped even one unit by December, unless they were planning to hand-build them with 3d-printed cases.

Also, look at the numbers. They had a $1.27m seed round in Feb '15, which lasted them until May of that year. In May '15 they raised $9M in series A, which lasted until Aug '16 (15 months) when they raised another $22M in Series B. They’ve been growing (up to 63 employees now), so assuming their burn rate has increased at least 50% since they ran out of the series A money, they’re possibly approaching the end of their runway before next spring, especially since their burn rate for series A money didn’t include mass production of Glowforge units, just R&D. They’ve got $28M or so in revenue they can start to recognize as they ship units, but a good chunk of that cost has to go to the cost of the actual unit. I estimate that they would need to ship 350-450 units/month to recognize enough revenue to keep afloat after their venture capital runs out. The only way I can think of that they could possibly do that is retail.

Let’s say they thought they could ship all units by mid-2016 when they ran the preorder. If they were only shooting for $2-5M in sales, even at a 50% profit margin on the units, that’s barely enough to cover 4 months of operations, max.

Of course there are lots and lots of assumptions here, and Dan is probably laughing about how far off my estimates are (I hope it’s because the situation is much better than I’ve surmised!) We’ll see what happens in the coming months… they may go for funding round C or be acquired.


#9

I believe @dan said they had funding for five years via backing.


#10

The quote from Tony was spot on - our internal target was $2M.