How (not) to sell PCBs: PCB drawer first year sales breakdown

in how-to by Ian | 49 comments

A while back we started selling Bus Pirate v3.5a-SOIC “DIY Edition” PCBs from the free PCB drawer for $3.95. Our informal survey and found that $5, including shipping, is a sweet spot for one board. So we hacked Zen Cart to accept PayPal for only one item, and started taking orders. With shipping, two boards from the PCB drawer are $9.90, right in the $5 zone.

Here’s a breakdown of the trial, based on our 2011 sales tax return.

100 PCBs were made for $124.90, including shipping. We contracted fulfillment to handle and ship orders. They charge $25 for receiving the stock and warehousing it. We also provided our own envelopes ($21.67 with shipping). The total startup cost was $172.

In four months from September to December 2011 we sold a paltry 40 boards to 29 customers for 158 bucks, $227.25 including shipping. PayPal charged us $16.12 in fees, 7% or about 55 cents per order, to receive the money. Initial take: $146.94 in PCBs, $64.40 in shipping.

US shipping is under priced at $2, international is over priced at $3.25:

  • A US order runs about $0.88 in postage, 22 cents for the envelope, and $1 for fulfillment ($2.10)
  • Non-US orders are $1.25 postage, $0.22 envelope, and $1.25 fulfillment ($2.72). Initially international shipping bounced around a bit, as high as $4 total, but it seems to be under control now

We received $64.40 in shipping and handling. Our aggregate fulfillment cost was $81.69 for 29 orders, plus $6.09 in envelopes ($87.78). We lost $23.38 from PCB profits on shipping differences. This could have been avoided by doing it ourselves, but even then someone should be compensated for handling the order.

From the 40 PCBs sold we received $146.94, less $23.38 in shipping overruns ($126.56, $3.16 per PCB). Each PCB costs $1.25 plus $0.25 warehousing, leaving a profit of $1.66 per PCB.

With $66.40 of profit we can’t cover our existing costs like the remaining PCBs ($75), envelopes ($12.60), and warehousing costs ($15). In the end we’re $36 bucks in the hole in this experiment.

Obviously this is a failure in the business sense, but profit was never a motive. We wanted to:

  • Test real orders in Zen Cart with PayPal integration
  • Learn about the cost and time involved in retailing PCBs
  • Judge the demand for $5 PCBs
  • Learn about our local sales tax licenses and returns
  • Get a bit of publicity by putting PCBs into the hands of dedicated builders
  • Try something cool

All of these goals were met, so we consider it an overall positive experience. Definitely not something to expand or continue beyond the first batch though.

There’s still a handful of PCBs left, get one while you can.

Grab the raw sales data spreadsheet from Google Docs.

This entry was posted in how-to and tagged , , .

Comments

  1. Arup says:

    Definitely, it was not a huge success.
    Oh, and what about the USB IR Toy PTH version PCB for sale in Seeed for $7+shipping? Was that flop too?

  2. Kevin says:

    This is interesting, thanks for the information. I have been tinkering with the idea of selling my own kits / boards for beginners. I will certainly take this into consideration.

  3. Tamur says:

    Interesting. The PCBs still left unsold should be counted as an asset. You can trickle them out eventually. Why would you guys want to contract the fulfillment of only a 100 PCBs? Could have saved that little bit of money with a good old diy :). That puts you right there in the profit zone even with such a small batch.

  4. Dolabra says:

    Thanks for the info.. I did my part! I bought one! I would have thought it would be a bigger success. Sigh.

  5. Cosmin says:

    This is publicity also, and cause I liked it I bought a BP4 and a FT2232H from Seeeds.
    but I like more to plant a pcb.
    if there were more different available for buying probably I would have bought more.

  6. RLW says:

    You might want to look up the definition of the word “profit”…

  7. John Burton says:

    I didn’t even know these were available… if I’d known I might have bought one instead of a completed one. I guess that doesn’t necessarily improve your profit though

  8. Sobh says:

    I was thinking why is it only one PCB was put on sale?
    Wouldn’t be better to put PCBs from other projects for sale to get wider base?
    As I would think (and I might be totally mistaken) that some of the initial cost would be the same, e.g. envelopes.

  9. Sobh says:

    Shit. No one told me that I can buy PCBs from the drawer thing. Are you going to not do so anymore?

  10. Sobh says:

    I didn’t know that I can buy PCBs from the drawer thing.
    Are you going to stop doing that?

  11. Sobh says:

    Ah, I double posted. Sorry for that :D

  12. Fulfillment overhead in a mail-order business is going to be a constant minimum number. YMMV, but I’ve found my fulfillment overhead to be around $7 per order. So on a single board that cost me $3 to get built, that I am selling for $10 I am making only $3.

    So even if you consider those boards scrap you still have to charge more than they cost you unless (as you say), you are doing it for the promotional value.

    PS) I sell my obsolete nixie driver boards this exact the same way – zen cart accepting only paypal.

  13. Jay Wilkinson says:

    I would have and still would buy more of the boards that you have for free. Right now I have 2 free boards inbound which I have most of the parts for already and already have plans to use my next 2 codes when the 2 that are inbound are complete and posted. I’d buy 2 or 3 more right now instead.

    Also I did buy the one above. From a cost perspective, it’s really not worth it to buy and build the board as money savings exercise. In fact once I factored in a Pickit3 I actually spent quite a bit more than to just buy it. And let’s not even count the FX-888 and other stuff I bought to build it. All told I probably spent over $200 to build a BP3.5! And I’d do it again! I’d still buy and build a BP4 and would also probably do the future versions too. I built mine because I’ve always wanted to try my hand at SMT and the BP seemed like the perfect project. Also yes I realize that I now have all the tools to build other things and now I always will.

    On one hand I don’t see your experiment as a failure. It wasn’t very clear on the site that only 1 board was for sale and the others weren’t. In fact I first tried to buy 3 boards and then after I emailed you and you explained it. I later bought the one. I think you should start adding in more and maybe all the boards for sale. I’d gladly PayPal you right now for a few more boards that I’ll end up getting codes for and building once the 2 inbound are done. Basically, I’d be willing to PAY to NOT WAIT. And yes I’d still post them and let you use them as usual on your site. It’s a way to make money off something that you’re going to give away anyway and still get the benefit of people posting and generating interest.

    Also I think you should consider kits. You may be able to work out with Seeed the purchasing of kits. Also an interesting option is selling the kits partially assembled. I just bought 2 of these http://rayshobby.net/blog/?page_id=1383 kits. He does the SMT himself (In a skillet (I wonder if they’ll smell like bacon)) and then sends out the kit and you install the through hole parts.

    Again, I don’t think you failed.

    • Jay Wilkinson says:

      On a side note, I’ve considered (in the case of the Part Ninja) ordering a batch of (your) boards from Seeed and building one or two for me and then mailing you the rest to give away. At least in theory they are open source hardware and I could do that. Of course I wouldn’t sell them but that one “sold out” in hours and just hours before I got a code. :(

      What are your thoughts on this?

    • Jay Wilkinson says:

      Didn’t see your previous post. Now there’s more boards I want to build. And I’d still buy them. ;)

  14. K Scharf says:

    Those fulfillment warehouses make use of what boils down to slave labor. There is an article in the current Mother Jones News on just how bad working in one of those places is. It’s enough to really make me think twice of buying stuff from Amazon.com anymore. If I every start an online retail business venture I’d rather do the shipping myself than give those warehouses my money, I’d sleep better at night.

  15. Mike Tsao says:

    The hangup I have with selling PCBs of an OSHW project is that the PCB is the smallest possible tangible representation of the work you’ve done designing, debugging, refining, promoting, and supporting the product, so it’s going to cut very close to the bone in terms of profit margin. If you sell an assembled or kit project, you’re getting the opportunity to mark up the components. It’s a fair trade because you, the seller of the kit, advanced the $$$ to buy components from Digikey or Mouser in full reels, and took the risk of ending up with 99 of a product nobody wants. But if you sell just the PCB, then there isn’t any reasonable way for the buyer to compensate you for your work. (I’m putting my OSHW community blinders on here and assuming that people are OK with compensating others for hard work, rather than the usual assumption that buyers are simply looking for the lowest price possible.)

    Charging a higher price for the PCB doesn’t make sense, either, because I’m a forgiving community buyer but not foolish. It wouldn’t really be rational for me to pay you $15 for a BP PCB, because then I’d pay another $15 or so in retail single-unit prices for components from Mouser, then another $7 for shipping them to me. I do get the DIY pride of using something I made, and that is worth something, but I feel like I’ve paid profit to the wrong people (Mouser and USPS) when I’d rather have that money go to the person who created the circuit.

    I’m not saying that bare-PCB sales are doomed. I’m just having trouble seeing how to get the money to move in the right direction.

  16. JesseJ says:

    I agree with the crowd. Love getting the data, and love seeing someone take a smart look at business practices. Jumping in and spending your whole bankroll would be a bad idea. Testing the market is critical. I also like the input from the other readers about possible modifications (like doing the SMT installs and ‘user kitting’ the thru hole stuff).

    I’m not sure how you arrive at the $15 estimate for warehousing (at $0.25 per board). Is this what your third party charges you?

    I once bought one of those little folding wire baskets that goes into a boiling pot of water. As I recall, the ‘price’ was at $25. I only wanted one, and no upgrades. The unit weighed about 8-10 oz and after shipping and handling it came to ~$50. I guess that “handling” can really getcha!

  17. AndThen says:

    15$ warehousing is the Annual end of year loss from the accounting viewpoint, 60 units at 0.25

    “As seen on TV” handling fees are always high, it’s where they make the money. Oxiclean is/was an exception, they pretty much don’t care what they’re selling as long as they get to add on that 10-15$$ handling fee.

    Postage fluctuation seems to have been very bad.

    Maybe try “media mail” for US post, I forget the exact discount. Ymmv, our counter clerks accept any media type that is not specifically parcels or letters. I’m sure i could convince them a PCB is a modern media, “See look silkscreen documentation”, I see little difference between a thumb drive and naked pcb, and expect they will also.

  18. msr says:

    These are bad news from my point of view. Open source (hardware or software) are cool, but if it doesn’t allow you to pay your bills… It would be much cooler if you can make enough profit to maintain an opensource activity for fulltime. IMO, open source hardware should not be “as cheap as possible”, instead it should be “good enough” to don’t be that cheap…

  19. Drone says:

    I didn’t even know you could buy the dang things.

    • Sleepwalker says:

      There’s a fair size flashing advert at the top right of every DP page, the Bus Pirate PCB is one of about 3 or 4 ads that come up.

  20. Lawrence says:

    I would have bought PCBs if you opened up more of the PCB Drawer. BPv3.5 isn’t that interesting as I already have one prebuilt from Seeed and don’t need another. I bet a bunch of other people are in the same boat.

    • Ian says:

      There’s a couple reasons we don’t sell the free PCBs, I should probably make a FAQ somewhere.

      1. PCBs are scattered all over the world on 3 continents where various DP community members are. When a PCB is chosen the person working on the board mails out the extra. If we sell those boards we get into a complicated international tax and business nexus situation that we can’t afford to figure out.

      2. They really are prototypes. They are the extras from in-progress stuff in our workshop. If we sell them invariably there will be buyers who expect support, firmware, software, partlists, etc that we have no ability to provide (and isn’t worth the $1-$2 profit).

      3. We’re really slow about mailing free PCBs, usually only one batch every 2 weeks (sometimes more!). We don’t want the responsibility of filling orders every day, which would be expected by paying customers. At only a buck or two a board it takes away from our actual profitable activities, but it’s ok as a free promo.

      4. Supplies are limited. We have no ability to send replacements when boards don’t show up, which paying customers would expect.

      5. If we priced it accurately (price of an engineer to stuff envelopes), most people would rightly say it’s cheaper to get 10 from Seeed or whatever.

      6. I have not had time to build the PCB vending machine/auto mailer robot yet :) It would have a live webcam so you can see your PCBs drop into the envelope and get labeled :)

      Many people pushed to buy free PCBs, which is why we figured it was worth burning $200 plus time on this experiment. It went ok, but I don’t foresee having a large stock of PCBs for sale any time soon. What we can do is increase the number of (somewhat) proven PCBs available at Seeed.

      • Sleepwalker says:

        I had also wondered why you didn’t sell the boards, but now you’ve pointed out all these things, what you say does make sense. I very nearly bought a bus pirate PCB, but when I roughly calculated the cost of the parts, it simply didn’t make sense compared to getting one ready to go from Seeed.

      • Jay Wilkinson says:

        So we need to design the PCB vending machine/auto mailer robot and help you fund it. How much space do you have for this? Can we assume 120VAC @ 20AMP is available? I’m guessing you’d object to a re-purposed cow tongue (commonly available from butchers) for licking the envelopes. Does the device need to make the board too? Or can we assume pre-made boards?

        I see this as totally doable.

      • Ian says:

        Premade boards. 240 is available too. Cow’s tongue is a tasty treat, not for licking machines! :)

      • Jay Wilkinson says:

        Wasn’t sure if you see the humor in that. :) I started a thread in the forum http://dangerousprototypes.com/forum/viewtopic.php?f=19&t=3724 it’ll be interesting to see if something becomes of it.

  21. adam says:

    I absolutely don’t get why the 100 boards would have been stocked elsewhere for $25. then extra money spent mailing out envelopes to the shipping company. seems like a silly way to keep costs low. I’d be surprissed if by doing thigns this was DID lead to great profits. not seeing common sense in this one. it’s really not that hard to keep 100 boards in a drawer at home, and just mail them out as needed. but maybe I’m missing something and misread.

    • Mike Tsao says:

      Ian did indirectly address that point when he said “This could have been avoided by doing it ourselves, but even then someone should be compensated for handling the order.” Unless the person you’re paying to fulfill orders happens to live in your house, then the boards and packing materials have to be kept somewhere else, and it costs money to get them there and keep them there. If you’re willing to DIY completely, Ian’s original point still stands: you need to put a value on your time and charge that against revenue, or else you’ll say “Yay! I’m profitable!” but wonder why your life sucks because all your personal time is spent taping boards to the inside of envelopes.

      Moreover, I see this as an experiment designed for scaling up. Assuming the concept is successful, can DP now do it for 100 different SKUs? It would be a more valid experiment the more they do it as if they already had 100 SKUs (i.e., not out of a closet at home). It would be a bummer to conclude you’re profitable at 10 units a month, then put your project on Kickstarter, get 2,000 orders, and realize it would take you years to fulfill all of them using the rinky-dink process that led you to believe you could do it profitably. (You can see this playing out with many Kickstarter projects at this very moment.)

  22. @Mike. Well said. I have done allot planning on how I will scale my board business and the most important thing I can say is track ALL your costs including your own time. Pay yourself like you would need to pay a skilled assembler/rework technician. Add $5 or $10 for fulfillment, and maintain a healthy profit margin after that.

    Right now here are my options for getting boards built:

    A) I pay a capable individual I met through a local hackerspace to build my boards. He gets paid per-working board. I mark up them up by 30% or so. I also factor in $5-10 to get the orders filled. I could also farm out the fulfillment job the same way pretty readily when I the time comes. This is what I am doing now, and it is OK for just starting out, but from a long-term business stand-point I should never expect to get ahead and grow the business. Plus, my job is to keep the design pipe full. Only I can do that right now. I need to reduce my cost of materials and assembly labor. Materials will go down by 1/3 to 1/2 once I am ordering parts in 100+ batches. Assembly cost is the area where costs can be reduced (See option E)

    B) Once I am doing a demonstrated volume of business (> 100 boards per month) I can try to convince someone like Spark-Fun to manufacture and sell my boards. They put up all the capital and risk of manufacture and in return, I would get a small, single-digit % royalty. They will also consider doing just the fulfillment. http://www.sparkfun.com/tutorials/315

    C) (@Ian, please correct me on this) I could also use Seeed Studio’s ‘Propogate’ Service, which is like Spark-fun except initially *I* must pony up the money for the parts (until I became a ‘strategic partner’ or > 1k pieces), Then Seed gets to do fulfullment. You end up with 10% or so after a number of things like handling and stocking fees are paid. http://www.seeedstudio.com/wiki/index.php?title=Propagate_Manual

    D) Use a traditional Contract Manufacturer, In low volumes (< 500 per batch) will cost more than the hacker-space guy, so margins shrink < 30%. Unhealthy if you have to share that 30% with retailers who will want a wholesale discount, who are the key to increased sales) Some CMs I have looked at – http://www.aapcb.com/, http://www.screamingcircuits.com/

    E) Setup up my own manufacturing line (paste screen jig, pick and place, reflow oven). Not counting parts or space, this would cost $30-50k or so for capital equipment alone. Probably more. The only good news about this approach is that my margins should climb up into the 90 – 120% range which is where they need to be for things to be healthy and sustainable. Further reading: http://www.ladyada.net/wiki/mdcpickandplace

    For me, E is the probable path. For others, especially those with day jobs, the royalty option may be a better route.

    ..c

  23. JesseJ says:

    Another point to also consider is “good will”. I know it sounds like altruism, but the small startup needs to get it’s name out there and build a reputation of ‘dangerous behavior’. I am just getting back into the DIY world, and am enjoying the hell out of the various stories of people getting their first tastes of DIY technology. I work in technology and have also been a teacher for over 30 years (yes, university level computer architecture and FPGA stuff), but the sheer enthusiasm of people MAKING STUFF and struggling to get it to work is fascinating. Enthusiam is contagious, and I’m still learning stuff along the way. If you’re not losing your shirt, keep it up! There is value in having a big audience, and whatever bait works is good! Follow your heart.

  24. Louis Mamakos says:

    I would echo some of the previous comments that I’d buy some of the other bare boards since I already have a bus pirate of my own..

    As a relative newcomer to the site, I confess that I’m a little confused by the marketing messages that I get from DP around the “free PCB drawer.” It feels like something that you communicate pretty strongly and on a regular basis – I see it at least twice a week via an RSS feed from the site and on Facebook. I’m excited about getting some of the boards to build — you’ve convinced me that I need some of these things, but then I can’t actually buy the thing from you. It seems like I have to enter a lottery that I’m unlikely to win for the the chance. It ends up being a slightly negative experience, and I’m really not sure what the point is?

    You reward people who build boards with the chance to get another one, but it seems unlikely that I’ll ever get a chance to bootstrap myself into the process by getting my first free board.

    I get that ideally there’s a revenue stream coming from an activity like this that’s profitable enough to make it worth the effort. As a customer, it would be cool to be able to get the board and a set of parts to build one one of the projects. Perhaps I get the board from you or someone else that as them stocked, and maybe a BOM for the parts I can easily order from Mouser or Digi-Key? Maybe the latter two could pay you a commission on the parts?

    I’d rather order a board from you and support the effort, but it seems like a more practical approach is to send gerbers off to some PCB prototype house since there’s no other obvious way to get them on demand from a supplier?

    • Sleepwalker says:

      I agree somewhat with what you say Louis, but having read Ian’s explanation, I can better understand why he doesn’t. Having said that, I did think much like you before and I still do to some extent, though I can also see his logic. Generally speaking, if you really want to make a board and are very keen, if you send Ian an email and explain and ask nicely, there’s a decent chance he’ll just send you a board. That would get you ‘into the loop’, so to speak, so if you then build it and post some pics and experiences, you’ll get to grab another. The BOM is usually in the details for the project and most parts are easy to obtain and/or have common supplier numbers, so all pretty easy.

      • Ian says:

        Thanks for your honest feedback Louis, I’m sorry it’s a negative experience for some people. We give away the PCBs because we have extra, and it’s a lot more fun to build boards with other people than by yourself. It’s financially, legally, and practically impossible to sell PCBs from the PCB drawer given the structure of this operation.

        As you note – please send our gerbers (available on each project wiki or in SVN) and have your own boards made. You’re even welcome to sell the extras (even in our forum). You can get 10 small boards delivered for as little at $15 with Seeed or Itead, etc, which is cheaper than I could sell you a single board if the PCB drawer was open for sales.

        Also as Sleepwalker said, unless there is a huge backlog of boards to mail or a flood of requests, we usually shoot a coupon out to anyone who asks. Especially anyone who followed project development in the forum.

      • John Burton says:

        Maybe have a forum thread for each board where people can express interest, and when say 8 people express an interest one of them gets a set of 10 made and sends them out to the others, and perhaps 2 for the free pcb draw?

        Probably too much of a nightmare for anyone to organise safely though I guess.

      • Arup says:

        That’s not a good idea John.
        If one out o 8 persons get full pack of PCBs and tries to send 7 of them to 7 locations. Then he’ll end up with more than $15 in postage, packing for the 7 PCBs. And at the same price he could have purchased the total set of 10PCBs from Seeed.

  25. John Burton says:

    Well I was more thinking that once 8 people expressed an interest they’d each send $5 by paypal or something to one person who would place the order and send out the boards once he had the money. It wouldn’t save people that much money but it seems such a waste ordering 10 boards when you only want one.

  26. John Burton says:

    But yeah, probably not workable.

  27. Jay Wilkinson says:

    Have you considered putting projects on http://batchpcb.com/index.php/Products and selling some of them there? This would be great for things like the TCPIP to Serial Converter where you mentioned “Yeah, the market is pretty saturated and that’s why we never really continued this project. As far as I know though, there are no similar adapters that are open hardware.”.

    Maybe not as a product but as a build-able board. Bus Pirate Demo board v4.0 would be another. You decided to go a different way but some people (like me) might still want one. Can’t wait until my board arrives!

    Just another suggestion for another way for you to make money, especially since you’ve poured time and money into these things that you may see as failures but that I, and probably many others, do not see as failures. Again just an idea.

    • Ian says:

      When we’re in shenzhen I’m going to talk to a group that plans to sell open boards and make part kits available as their profit generator. That could be similar. You can always put our gerbers into batch PCB yourself, it would be the same cost I think(?).

      • Jay Wilkinson says:

        I would consider using them and doing that to make a single board of one that you don’t have in stock, like the Part Ninja. But I’d never upload them and set them for sale. Anything like that, that would generate a profit off your work should only be done by you in my opinion.

        I’ve also considered sending the files on the Part Ninja to Seeed and having a set of 5 made and then mailing the remainder to you to give away or posting them in the forum for free. Maybe I could trade them to other people for unused codes and order more of the boards that you do have in stock. :)

        P.S. My boards arrived today so I should have more projects posted soon.

Leave a Comment

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Notify me of followup comments via e-mail. You can also subscribe without commenting.