Will Your Hardware Startup Make Money?
medium.com
medium.com
(1) This is a decent estimate for a completely commodity product (Bluetooth headphones). No iOS/Android app development, no content development, very little custom firmware or electronics development required. (That correspondingly means you face the least differentiation and the most competition.)
(2) This is for a "plastic and PCB" product build. That's as simple as it gets. Any product that goes beyond a PCBA inside a plastic enclosure is usually going to have significantly greater design and engineering challenges, not to mention COGS costs. Moving parts? Temperature? Water? Cameras/lenses? Sensors? etc.
(3) The marketing / customer acquisition cost is either ignored or assumed to be some laughably small number. Even at its highest, this article assumes $3/unit customer acquisition cost. That's ridiculous (go look at Fitbit or Gopro or any other filing statements). If it were true, a lot more people would go ahead and get that initial capital to make their own Bolt-o-Phones.
(Disclosure: co-founder of a hardware startup https://www.pantelligent.com/ and about to ship our first production run.)
i'll look for them and update this comment if i find them
----------
> We have recently begun to spend significant amounts on advertising and other marketing campaigns to acquire new users, which may not be successful or cost-effective.
> We have recently begun to spend significant amounts on advertising and other marketing campaigns, such as television, cinema, print advertising, and social media, as well as increased promotional activities, to acquire new users and we expect our marketing expenses to increase in the future as we continue to spend significant amounts to acquire new users and increase awareness of our products and services. In 2014 and for the three months ended March 31, 2015, advertising expenses were $71.9 million and $21.1 million, respectively, representing approximately 10% and 6% of our revenue, respectively. While we seek to structure our advertising campaigns in the manner that we believe is most likely to encourage people to use our products and services, we may fail to identify advertising opportunities that satisfy our anticipated return on advertising spend as we scale our investments in marketing, accurately predict user acquisition, or fully understand or estimate the conditions and behaviors that drive user behavior. If for any reason any of our advertising campaigns prove less successful than anticipated in attracting new users, we may not be able to recover our advertising spend, and our rate of user acquisition may fail to meet market expectations, either of which could have an adverse effect on our business. There can be no assurance that our advertising and other marketing efforts will result in increased sales of our products and services.
ALSO
> Revenue increased $474.3 million, or 175%, from $271.1 million for 2013 to $745.4 million for 2014. A substantial majority of the increase was due to an increase in the number of devices sold from 4.5 million in 2013 to 10.9 million in 2014, including $151.9 million from new products that we began selling in 2014. U.S. revenue, based on ship-to destinations, increased $356.5 million, or 173%, from $206.1 million for 2013 to $562.6 million for 2014 and international revenue, based on ship-to destinations, increased by $117.9 million, or 181%, from $65.0 million for 2013 to $182.9 million for 2014.
So ballpark 11 million units sold in 2014, and 71.9 million dollars in advertising in 2014
That's $6.59 worst case customer acquisition cost (assuming no repeat customers)
http://www.sec.gov/Archives/edgar/data/1447599/0001193125151...
In contrast, a hardware startup (or a software one, for that matter) just has to go out and hustle to get every single sale, whether through paid or non-paid customer acquisition channels. This is where YC's philosophy of treating hardware and software companies similarly works: they both usually have very similar customer acquisition challenges in rapidly building and scaling their growth channels (rather than, say, technical engineering challenges, which are very rarely the limiting factor in startups).
Kickstarter, obviously, is a big one. Never mind what people say about needing your own audience - yes, but you gotta start somewhere, and if you've got the message, Kickstarter's got the soapbox.
Another is simply being media-saavy. The press has to write about something, so don't forget that they need you, too. But you have to learn how to pitch stories.
If you had an original product that actually goes viral, the marketing costs goes way down.
Kickstarter has really changed the game. These guys launched today, for example, and they'll probably be funded within 24 hours:
https://www.kickstarter.com/projects/1152958674/the-sensel-m...
I might just suck at PR, but I don't think The Sensel Morph is really representative of your run-of-the-mill funding experience.
https://www.kickstarter.com/projects/keyboardio/the-model-01...
I think a lot of these guys try to build a following before they launch.
I might just suck at PR
You just picked a really hard product to make successful.
The problem is J. Random Gamer has no reason to buy it, unless enough of his gaming buddies have it that he gets shut out of games if he doesn't have it. If he's the only person in the game who has it, or even if half the people in the game have it, the cheaters are still in the game and he's spent $125 without any benefit. And that's assuming the games support it - if no games support it properly, there's even less point.
There's a chapter about this sort of thing in Thomas Schelling's Micromotives and Macrobehaviour, if you're interested. Not cheaters in pc gaming (the book was written in 1978) but the game theory around similar situations.
One option is to make the product known to people who run the competitions, then wait for a major cheating scandal and lobby them to make it compulsory - but obviously that would only give you pro sales, not the much higher volume consumer sales. It might be possible to arrange a bunch of great servers where the device is mandatory, so early adopters get some benefits.
In the longer term I'd also suggest some sort of IP licensing, where something compatible with your product is built right into gaming mice. That way users have one less thing to cart around, and companies who know how to make and distribute hardware can take care of that for you. And you'll have a much easier time getting it made mandatory if it's "$1 in every gaming mouse from a variety of competing manufacturers" rather than $125 with a single source :)
Interesting product... they are almost there and the traffic from here should do it.
Yes, you will likely fail if you've never done this before and you're trying to make a consumer-quality product. So get experience with an intermediate step or two.
But it is certainly possible to make money on such a first batch: Why do you need inhouse EE, ME, DE en Embedded Engeneer for a headphone. Many of these functions can be done by a small design firm or freelancer that you hire for a much smaller fee. In addition, these people do this all the time for different companies and have a huge experience this way and good contats (you can also go with teh bigger firms, but then you probably need the same amount of money).
I think you really need some expertise in what you are trying to build. It has to be really hard doing a decent cost estimation otherwise.
- http://www.pcgamer.com/introducing-gameref-the-anti-cheat-ha...
- http://www.redbull.com/en/esports/stories/1331720277392/game...
- http://www.pcgamesn.com/gameref-is-a-prototype-anti-cheat-sy...
- http://www.polygon.com/2015/2/23/8090221/onling-gaming-cheat...
- http://motherboard.vice.com/read/can-this-little-device-fix-...
.. it failed to get funded. Failure is part of the game and I wasn't too phased by it, but it still sucks. Unlike a software start-up where I can build and launch a working alpha/beta, you NEED some money to pour into a hardware start-up right at the start. It's not easy and, as the article mentions, the margins aren't as great as software. The sentiment of hardware being tougher than software is ubiquitous from what I can tell.
Props to the people that make it though.
See the KS page here: https://www.kickstarter.com/projects/1094040691/game-ref-the...
My goal was $200k. Some people criticized it for being too high, but I think it was actually a pretty conservative estimate. Oh well, c'est la vie.
Is there the same transferability with hardware startups? Do other companies love to hire failed hardware entrepreneurs the way they usually like to hire failed software entrepreneurs, or do you just write off the money and time spent on the startup as a dream forgotten?
This isn't my first rodeo, but I usually just write off failed start-ups in general as dreams forgotten (and boy do I have a lot of forgotten dreams). I obviously gained a lot of technical knowledge: how USB works, packet formats, some hardware engineering, driver programming, etc. But unless I ever get a job writing low-level code for hardware devices, I doubt I'll ever use that knowledge again.
Who knows, maybe for another hardware start-up?
I believe hardware anti-cheats will be more efficient but until we have companies (cevo, esl, mlg etc) that are willing to front the cost of the devices for the customers, it'll be a short lived thing.
I'd imagine, some form of a rental system would work market fit wise, sort of like a DVR from your cable company.
Until then, software seems to be the easiest solution sadly.
If my understanding is correct, product dev teams seem to aspire to an initial 500 - 2000 units for that initial run as the symbolic point of reaching this "finalized" level of development. Any production at this level is considered a finished good ready for the sales/supplier pipeline.
Once a product has reached this level, my understanding is that subsequent runs is typically for 10K to 25K widgets per run ("lot") that see anywhere from 10,000 to 250,000+ widgets for a single SKU per production run.
Above 1 million units, everything changes, but the 5K first-and-only run seems abnormal.
Also people often forget about your suppliers management, accounting, and sales overhead. That means that they will have minimum orders below which they don't make enough money justify the overhead of shepherding your product through their shop. For Chinese suppliers 5000 units appears to be typical because lower volumes are typically on-shored. On shoring is going to increase your costs by probably 30%.
You might ask the other question, why not 10,000 or 20,000 units? Answer is that starts to be a huge chunk of money for something that isn't proven.
So ans: 5000 because less the NRE kills you. And you'll have to on shore production. And no more than that because otherwise your initial production costs becomes the dominant part of the investment risk.
Thanks for the post.
PS: Full Disclosure, i launched a robotic haptic device (cubify.com/touch).
How did Cubify handle these things?
To sell 50k units or more you will need a few people to handle your store, marketing and support. Also will you continue to sell the same model for years or develop new versions? Lots of other costs to keep a business running should be included in your business financials.
A 3x multiplier is slightly low for CPGs, I'd go with 4-5x depending on the product and market to give yourself some breathing room. Things like sales, resellers and coupons all eat into profit.
Also, why is the cost of employees with overhead lower? (NB/EDIT: Pointed out below this is for 3/4 of the year, not the full year - but the rest of my point stands.) That makes zero sense. Surely they are paying payroll tax, unemployment insurance and health insurance for their employees, no?
I think that's the cost of the employees for the 9 months (not 12) it takes them to do the required work.
Can Protomold in Minneasota rotoshill my parts?
I don't know what rotoshilling is but manage to produce hardware anyway. A problem I see is that first-timers think they have to act like the big boys, and the big boys design complex assemblies and go to China. (I wouldn't unless you have the money to either make an expensive mistake, or to hire someone with experience.)
Long story short on all this, if you've never made hardware before and are going to make hardware, take the risk with other peoples money... if at all.
I've heard the fully-loaded cost of an employee is about 2x their salary, but I'm interested to know what the fully-loaded cost of a startup employee is. At the minimum, an employer would probably need to pay for health insurance, for example. Even if the benefits aren't very good, the cost is higher than someone's salary.
The article says the cost with overhead is 1.25x, but is that realistic?
Think of:
- Health Insurances ($500/mo)?
- The other half of payroll taxes (~%8)
- Food benefits ($10/day in snacks, $15/meal served)
- Office rental & upkeep costs
- 401k program costs (yes, they charge employers for this)
- HR Costs
I know for a fact finding talented EE people is possible at $75k, and that same level talent would most likely cost $100k for a software dev. Interestingly, there is a decent supply of EE talent out there, however due the varying degree of hardware applications finding the right person is more difficult than it seems on the surface.
One of the best advice my Dad ever gave me when I was thinking about majors in college was that you could program computers with an EE or a CS degree but without an EE degree they probably wouldn't let you build/design them. The math and physics were, to my taste, more difficult but I got tremendous satisfaction in being able to understand a computer from the PN junction of its transistors, to the process model of its operating system.
My parent spout a lot of the same nonsense. Just who exactly are these "they" gatekeepers the Boomers keep referring to? Pretty sure "they" are the same keepers of great wisdom such as "Housing prices will always rise", and "Work hard, you'll get a pension and gold watch," sorts of people. Not intentionally wrong, but definitely behind the times.
Want to make a computer? WE don't need "them" to gatekeep their own bullshit anymore. Everthing, including fabs has been virtualized and is available online. Funding, PCB design and layout, soft-hard your FPGA ASIC designs, etc, etc, etc. Make your tradeoffs same as the rest of us, but don't think for a second that you must have "them" to make that computer.
I know it's not really that easy (my parents spout the same nonsense all the time, but I love them nonetheless), but try to surround yourself with enablers not gatekeepers. And be able to quickly differentiate between the timeless and the boomer knowledge.
if your orders double from 20K units to 40K, your costs will roughly double, usually meaning that any profit made on the 20K is re-invested to make the 40K - usually plus additional capital.
So long as your production quantities double, you will not actually "have" any money (unless your FOB cost is north of four times sell price). Many companies cannot "realise" profit until orders stabilise.
Hope someone can do an explanation like this for a software startup too.
COGS for us is how much is sold through a distributor to an end user.
You won't always be able to hire whoever you want/need if you offer them equity. An inexperienced, younger person who's willing to work for equity might cause delays in delivering which will also cost money.
1. http://www.forbes.com/sites/neilpatel/2015/01/16/90-of-start...
Consumer hardware is easy to purchase with a stolen credit card and fence on eBay.
On a serious note, it surprised me that the BOM listed the shipping materials as being about as expensive as an assembled and flashed PCB. It's interesting that a cardboard box (which feels pedestrian) costs about as much as a printed circuit board (which feels a little like magic).
I personally don't think this is a good case of Betteridge's by the way.
The subject at hand is a hardware kickstarter. That bridge has long been crossed and burned. If you had all the talent and resources at hand already, you'd probably have funding too.