Had a nice, abbreviated Christmas in Arizona. I got a model rocket that I intend to do something cool with, like integrate a sparkfun module or something. Speaking of Sparkfun, how crazy is this Free Day thing? I’m excited to see how that all goes!
For various reasons related to the out-of-stock problem I got to do my own laser etching on Logic for the first time in a long while. And this is by far the most I had ever done, so it needed to be done right (I did have an entire 2.5 hour window at Techshop, but experience has shown that things never go smoothly with the laser). I did a bunch of prep work in illustrator creating all the artwork. I had to design up a little cardboard (only thing I had around the office that would work) jig. Overall it worked great, although it did smell strongly of burned cardboard.
My brother is out here over the winter break and we are hammering out the software 80hrs/week. The reason for the latest delay – with the analyzers – is that I realized about a month ago the entire data architecture would have to have a very significant overhaul in order to be future proof, and EVERYTHING depended on this path and what we had better do it now rather than later. This is largely completed, and now we’re more or less porting everything over to use the new data layer. It is necessarily complex, and we iterated quite a few times to come up with something that didn’t explode in complexity and edge cases once you tried to implement it. I think we have something really solid now, and most of the scary (i.e. I don’t know if this problem has a realistic solution) and mission-critical problems have been solved. Frankly the fact that this is really hard stuff, some of it, certainly keeps it interesting. I readily admit that on the surface a Logic Analyzer does appear very simple. It is simple, to a point. But if you want to do really cool stuff, and do it really fast, it stops being simple fairly quickly. Take drawing to the screen as the obvious example. On the one had this couldn’t be easier. Is the input high? Draw a line at the top, and so on. But what if you have 100M samples on the screen at once? Now every pixel represents on the order of 500K samples. You need to know if all these samples are one, or zero, or something else. That’s not necessarily easy to do quickly, particularly in the worst cases. That’s pretty much why the current Logic Software (windows only) can really only handle about 200M or so samples before it starts to chug badly when zoomed way out. It does some fancy stuff to make this much better than it would be otherwise, but ultimately we want it to be lightning fast at all zoom levels.
I’m liking Windows 7. OS X needs to deliver something really nice in the next release, we’ll see what they pull off. The iPhone however appears to really have a nearly unbeatable position, thanks largely to the apps. I recommend it hands down to everyone, if you can stomach the monthly payments. My brother has the Pre, which is pretty awesome. I haven’t played much with Android. I’ll bet Apple is steamed about everyone mentioning how the iPhone can’t multitask. It’s almost the only area where they are venerable, and it’s one of those things that most people don’t even know what it means, but that doesn’t mean it’s not hurting sales. If I were them I’d give in and do some sort of compromise – allow memory and processor constrained “mini processes” or something like that. There are some cool things you can do with that – such as make a pedometer app, or a GPS logging app. But they do have a legitimate concern, keeping high performance on a phone isn’t easy to do when managing a ton of virtual memory, etc.
Little milestone over here – UPS has a sales guy that now comes by and makes sure we’re happy and provides some pretty good discounts, over and above the ones you get just by having a daily pickup. I keep pestering those guys to get me some UPS posters for the office, and it paid off, now I have a little UPS cast iron truck model. I really like UPS, (and container boats) because they literally represent ‘shipping’ a product, which is kind of what it’s all about if you like entrepreneurship. By the way, if you’re really interested in entrepreneurship, there’s some really good talks by entrepreneurs on iTunesU, largely associated with Stanford. It’s a little bit hit or miss, but there are some serious gems in there.
One of my personal heroes is Elon Musk (I’m getting a Tesla Roadster to celebrate ‘making it’ when that ultimately happens) and SpaceX is just about ready to launch their big Falcon 9 rocket for the first time. I think they’re a little behind, and for sure the pressure is on. Talk about stress, and a hard core problem. Makes me less stressed about the comparatively minor fires that are constantly being put out over here.
We pulled all our inventory back in-house, and that’s where it’s going to stay. It’s too messy to try to maintain separate inventory at your assembler’s location. To the extent we need assembly help, we’re going to just kit it from now on. We also got another office/space at Activspace, so no I have a fridge (with ice, and iced tea, soda, etc) a microwave, and tons of these cool storage shelves filled with inventory.
Speaking of inventory, we continue to have some issues there. This was mostly caused by the Logic Aluminum case debacle, which has ripple effects. But here’s my new and improved philosophy, which is somewhat contradictory.
1. Finished goods inventory is king.
2. It doesn’t make sense to have uneven quantities of parts inventory because parts that can’t be assembled into finished goods inventory is just tied up cash.
3. It makes sense to have tons of inventory of the inexpensive components because this 1) doesn’t require substantial cash outlay, and 2) reduces to a handful the number of parts you have to actively and carefully manage.
USPS is bleeding money, I wonder what they’re going to do to fix that. Not exactly an easy problem... I don’t really like USPS (understatement), but still they are tolerable and for cheap international shipping they are the only option. That said, I hope to be offering UPS as well, internationally, very soon (just takes some IT work), as our costs for that are coming down due to volume discounts and it’s really the only way to get something oversees quickly, and I think some of our customers (typically businesses) would rather do that than risk the open ended customs delay.
All for now, back to coding!
Yet again Saleae stumbles in the inventory department. One of our suppliers (not the one in the title of this post) who makes the Logic aluminum cases failed to deliver despite massive lead time. Ultimately I'm the one at the helm of this thing so it's my fault. And certainly my problem! Here's some takeaways:
1. If a supplier has historically been late and otherwise problematic, that's not going to change with time. (maybe it will, but hoping is not a good strategy)
2. Order product from your backup suppler before it becomes obvious that you're going to go out of stock. (duuh, I know)
So I called up our first supplier (well, the first one who did serious quantities) and luckily for us they were happy to help us out. It costs much more of course but at least we have parts coming in now. Meanwhile we're quoting/sampling with another, lower cost supplier that I trust from the other stuff they've done for us. AP knows this.
Check this out. The bar stock at the left is what Logic is made of. It's extruded 6061 aluminum. The shiny machined piece is a LOGIC-TOP part, completely machined on one side. It's not sitting on a square piece of aluminum, it's all one piece. This gets turned over and the back side is machined. Then the part is "tumble de-burred" to clean it up and give it a smooth texture. Then they go off to the anodizer, and last they are laser etched.
When I was down at American Prototype I shot this video. You can't see much, but you get the idea. It was extreemly cool to see in person. I really like the automatic tool changes in particular.
Cross platform Logic software released; Linux and OS X users rejoice; Deep sigh of relief had by owner
Yesterday we launched the alpha version of the cross-platform software. (private beta) I wore my new Ubuntu t-shirt. It went fairly smoothly. On Tuesday night we sent out an email about 10% of the signups, and uncovered and resolved one mission critical bug on OSX. Since then it seems to be working alright for people. The release was very feature sparse. In fact other than collect and display data, and give parameters like width & frequency, it doesn’t do a whole lot. But not to worry. The hard part is out of the way now (I’m SO happy about that let me tell you, holy crap). The beta version (1+ month out?) will have all the features that the current windows-only version has. And then we at Saleae can sit back and just let the money roll in. Not really =) Making some solid software that both works well and looks good cross platform is not particularly a walk in the park in my experience. Actually, knowing what I know now it’s not all that bad. But if it’s the first time… ouch. I think that the result so far is pretty solid as these things go. The beta version will be top notch. Sr. software engineer quality on the inside.
If you haven’t noticed, we’re now up and running in the EU, selling directly to our customers over there. This actually has gone really well, with minimal issues. The IT stuff is pretty ironed out, just a few more things would make it a bit better. Order fulfillment automation is really effective for the first 95% of things you want it to do. Actually automating the last 5% in my view would take x2 as much effort as the first 95%.
Inventory is okay, although we’re cutting it pretty close right now. We’ll be building a massive number of these things within the next month.
My brother came out for about 3 weeks and in addition to putting in 80hrs a week into the company we also managed to get our certification for Basic Keelboat, woo! The school is OCSC at the Berkeley Marina. Highly recommended. It goes Basic Keelboat, then Basic Cruising, the Bareboat Cruising. The latter of which means you can go to the Bahamas and charter yourself a nice 50ft catamaran if you’re so inclined.
We managed to scrap together enough loose change to afford an intern, John. He’s still in school but will be giving us a hand with some odds and ends. Some good stuff is in the pipeline! No new product, but we should be doing some cool stuff with the website in the not too distant future.
A few weeks ago we got another suite here (these suites are tiny, btw) and it’s dedicated to shipping. I can package up a Logic in 15 seconds. =) Actually I haven’t timed it.
I can’t believe it’s already towards the end of August. At some point your hopes and dreams get dragged back a bit by reality. But not to worry, I haven’t lost the spark =)
Thank you to all the heroes out there for trusting Logic with your debugging problem! As always, drop a line if you have a suggestion.
What weights 25lbs and costs significantly more than my car? Why, this box of Logic PCBs
Tool of the trade.
Finally, light at the end of the tunnel! We recently launched the cross-platform SDK and although it didn’t work perfectly right away I think it’s actually pretty close. Meanwhile the cross-platform app continues to come along nicely. I’m shooting for a bare-bones GUI release pretty soon. That said, I’ve been wrong about every prediction I’ve made with this effort so the new deadline is “it’ll be ready when it’s ready”.
Late last month we moved into a new space managed by a company called Activespace. So far I really like it. I sit right next to a bright window. The new address is
3150 18th St Suite 344
San Francisco, CA 94110.
It’s literally 4 blocks from the old place. Oh, the new place actually has parking (parking tickets were really killing me for a while) although it’s $.75/hr – at least it’s there when you need it.
While no one is buying a new car, sales have stabilized to a rate that definitely keeps the lights on, so that’s fantastic. Looking at the total number of units sold, it’s actually pretty mind boggling to have gotten to this point. Thanks everyone!
Inventory has continued to be the bane of my existence. Here’s the deal. When you run out of inventory, sales go down. And sure, you recover some, but overall you probably loose at least half the sales you would have made. Now inventory is hard to manage because it takes so much cash to finance – and when cash is tight and you have to choose between making payments on overdue invoices and inventory purchases… By the way, this is not at all an intractable problem. Our inventory could be managed like a champ if I had the time to put a proper system in place. Having plenty of cash generally makes managing the inventory much easier and we’re moving in that direction.
Speaking of putting a decent system in place, we have done that for order fulfillment. In January we started out with this fulfillment house called Webgistix. While this wasn’t a complete disaster, they were a major pain in the butt to deal with, their IT systems were apparently implemented by an intern, they would duplicate orders, lose packages, mis-count inventory etc. (Maybe we’re just high maintenance) This plus our declining sales made me decide about a month ago to pull fulfillment back in house. During that time and soon after our fulfillment infrastructure got pretty good (lots of php/mysql work). Now it’s just a series of button clicks to print invoices, shipping labels (including all the customs stuff), and send out various order confirmation/order shipped emails. The only thing that isn’t automated is anything that doesn’t go through the normal website sales channel. Although I imagine that won’t be too far off either.
By the way, about this whole Linux/OSX thing. I don’t think it was really the best business decision to have made in hindsight. On the plus side, it’s going to be awesome and we will definitely get some new customers (what % is a good question). On the down side (in hindsight) it has turned out to be an extremely challenging engineering undertaking, far beyond what I had originally imagined. It’s allowed our competitors in large part to catch up and reclaim their market share. That all said, and with due respect to sunk cost, I’m actually fairly happy to be doing it: 1) It’s going to be awesome and a nice feather in our cap 2) It’s been an fantastic learning experience for me 3) It’s given us a chance to re-architect the entire application from the ground up. And for better or worse, I’m more motivated by a vision of how awesome we can make something than with strictly business decisions.
One part of many in the Saleae world-class fulfillment implementation
The USPS API doesn't actually let you pay via the api, thus the make-your-own-stamps
Woo! Latest n' greatest
Ubuntu Linux! Running native on 2.0GHz core2 duo/ 3GB ram
A typical hour's worth of orders. Just kidding...
Saleae gets a mascot
We've switched from saleae.vox.com to blog.saleae.com so that it's easier to leave comments. All the past posts can be found here: http://saleae.vox.com/