How We Spent $500 on Tech to Ship $2.6M of Soylent • Hacker News Board

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#61

Post by geographomics » Thu, Jun 04, 2015, 2:51 PM UTC

yen223 wrote: post #9

scriptman wrote: post #5

Earlier quoted context omitted.

If you put an apostrophe in front of the number it will store it as text but will appear as a number. Just to be clear - you can then save the xls/xlsx to csv and the csv won't contain the leading apostrophe, but will show the whole number.

If you put the apostrophe in front of a number, and save it as csv, that works. But if you reopen that exact same file, Excel will still misinterpret the number. The issue here was that there's no way to let Excel know that this value in this CSV file is not a number. The only way around it that I know of is to stick with xlsx, which has its own pain points.

The way around it is, rather than opening the CSV file directly, to create a new blank workbook and import the CSV file using Data --> Get External Data --> From Text. This opens the Text Import Wizard and from there one can define the datatype for each column - set it to Text (instead of the default, General) and it won't be interpreted as a number.

Another way is to rename the .csv file to be .txt instead, and load it in normally. This triggers the Text Import Wizard upon opening, and the rest is as above.

Also to do this programmatically, one can use the VBA function Workbook.OpenText, and specify the data types via the FieldInfo parameter.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#62

Post by TeMPOraL » Thu, Jun 04, 2015, 2:53 PM UTC

vans wrote: post #25

Nice article about an awful product. I don't eat to live: I live to eat.

To each it's own. I like to enjoy a good meal every now and then, but day-to-day eating is mostly an annoying necessity. I have better things to do with my time.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#63

Post by TeMPOraL » Thu, Jun 04, 2015, 2:55 PM UTC

pavlov wrote: post #14

SQL2219 wrote: post #13

Don't eat soy.

Apparently Soylent doesn't contain much of either soy or lentils. It's a weird way to name a product. Sort of like making a fruit juice called "BeefPork". (I know the name is a movie reference. Still doesn't make sense to me.)

> I know the name is a movie reference. Still doesn't make sense to me.

It should, since you recognize the reference, and the name fits perfectly. For me, Soylent Green pretty much the first thing that comes to mind when thinking about this type of product.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#64

Post by frandroid » Thu, Jun 04, 2015, 3:16 PM UTC

virmundi wrote: post #29

harkain wrote: post #23

Earlier quoted context omitted.

Why?

There are various reasons, some of which will require time to checkout. One is that soy is a phytoestrogen [1]. So the studies show there could be some issues with men eating soy at high levels. Googling provides a lot of arguments founded in reasoning for and against. Another issue is that soy requires heavy agriculture and processing [2]. This can harm the environment in much the same macro ways as beef production.…

Considering that soy is being fed to cattle, by definition, soy cannot harm the environment in the "same" macro ways as beef production. Beef production is an order of magnitude more, and anyone who substitutes soy for beef is doing a favour to the environment.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#65

Post by protomyth » Thu, Jun 04, 2015, 3:47 PM UTC

TeMPOraL wrote: post #59

protomyth wrote: post #56

Earlier quoted context omitted.

You edit the data type because it isn't a number your importing, its a shipping tag. On import you set the column to text, its right there in the import dialog.

ID handling 101 - just because it's made only out of digits, doesn't mean it's a number.

How true that is, dealing with trains, barges, and government issued ids will teach some seriously valuable lessons on that score.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#66

Post by nols » Thu, Jun 04, 2015, 3:47 PM UTC

angersock wrote: post #42

chrisBob wrote: post #33

Earlier quoted context omitted.

I love math where $500 in hardware plus 6 months of engineer salary equals $500. I am a grad student making about $30k/yr and my advisor still says to figure labor as costing $1,000/day.

Startup math is a bit different. For example, without that kludge, they wouldn't have been fulfilling orders. So, even if they "wasted" half a man-year, they still made their sales and avoided the cost of, you know, going out of business. You can't really apply operational reasoning to startups the same way you can to larger businesses--you've got to factor in other things. (And yes, that is used frequently as a way…

That's not the criticism. They're not cobbling a solution together to prevent bankruptcy, they're cobbling together a solution to an already solved problem. Ecommerce fulfillment companies exist, and they chose a freight company with inadequate technology for their needs. Then they're justifying it by saying they hacked a solution for only $500, when it's really costing $500 + 6 months salary + hurt reputation for an extremely late delivery.

They could've incorporated a company that used EDI (or an API) to tie into their database to process orders much more effectively than hacking together a solution with a store bought laptop.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#67

Post by galago » Thu, Jun 04, 2015, 3:51 PM UTC

nbouscal wrote: post #21

"a $500 system that successfully managed the first four months of Soylent shipments." Successfully is rather overstating it. The shipment of Soylent was delayed by over a year and went at a snail's pace. My shipment was delayed at a rate of roughly one month per month, with it eventually arriving something like six months after the first shipment had gone out, despite being one of the first backers. I can't think of…

Very true. I didn't receive my shipment in a timely manner. I asked for and received a refund. In my case they left money on the table.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#68

Post by protomyth » Thu, Jun 04, 2015, 4:01 PM UTC

vinceguidry wrote: post #60

protomyth wrote: post #51

"A UPS tracking number is 15 digits or something like that, but a FedEx tracking number is 24 digits. If you opened it in Excel it would truncate it, because it would be like, “That is a 24-digit-number. We should store this in scientific notation.”" I am stunned that this is actually an issue even when dealing with CSV and Excel. It is quite easy to store numbers in Excel and treat them as simple text.

Excel really is horrible at handling anything that needs to feed into a software system that isn't also built on Microsoft technology. Just yesterday I was trying to get a coworker to manually order some rows in a database that I had exported to CSV by date, the problem was that some of the dates were just month-year, and some were month-year-day. Excel insisted on formatting the columns as dates and the month-year r…

Well, a lot of people use Excel to prep data for other systems and it really isn't that hard to handle. Perl has excellent libraries for reading and writing Excel spreadsheets. Excel is really easy with a large amount of training material. Even the Perl libraries are very easy to use.

I would never tell a user to open up a text editor on tabular data. It is too easy to configure Excel to do the right thing each time. Basic spreadsheet formatting takes care of a lot of problems.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#69

Post by angersock » Thu, Jun 04, 2015, 4:09 PM UTC

nols wrote: post #66

angersock wrote: post #42

Earlier quoted context omitted.

Startup math is a bit different. For example, without that kludge, they wouldn't have been fulfilling orders. So, even if they "wasted" half a man-year, they still made their sales and avoided the cost of, you know, going out of business. You can't really apply operational reasoning to startups the same way you can to larger businesses--you've got to factor in other things. (And yes, that is used frequently as a way…

That's not the criticism. They're not cobbling a solution together to prevent bankruptcy, they're cobbling together a solution to an already solved problem. Ecommerce fulfillment companies exist, and they chose a freight company with inadequate technology for their needs. Then they're justifying it by saying they hacked a solution for only $500, when it's really costing $500 + 6 months salary + hurt reputation for an…

( http://en.wikipedia.org/wiki/Soylent_%28drink%29 )

Look at the timeline. They only started shipping in like May of '14. This article covers April of '14. They had only run a crowdfunding campaign like less than a year prior. They raised funding (a lot) a year later.

The problem with "but but but they should've used an API or an order-fulfillment company" ignores that, at that time, they very likely were lacking the upfront capital required to setup and guarantee that relationship. The problem with "but but but they wasted a ~salary~" is that, if you looked at their payroll at that time, I'm willing to bet they were being compensated mostly in equity and not market wages.

Again, there is an entire solution space that is strictly better that is also strictly off-limits if you don't have the spare funding.

EDIT:

More to the point, this thread is full of Monday-morning quarterbacking. The fact is, these guys raised $20M, and we're bitching on the internet.

Re: How We Spent $500 on Tech to Ship $2.6M of Soylent

#70

Post by chrisBob » Thu, Jun 04, 2015, 4:45 PM UTC

angersock wrote: post #69

nols wrote: post #66

Earlier quoted context omitted.

That's not the criticism. They're not cobbling a solution together to prevent bankruptcy, they're cobbling together a solution to an already solved problem. Ecommerce fulfillment companies exist, and they chose a freight company with inadequate technology for their needs. Then they're justifying it by saying they hacked a solution for only $500, when it's really costing $500 + 6 months salary + hurt reputation for an…

( http://en.wikipedia.org/wiki/Soylent_%28drink%29 ) Look at the timeline. They only started shipping in like May of '14. This article covers April of '14. They had only run a crowdfunding campaign like less than a year prior. They raised funding (a lot) a year later. The problem with "but but but they should've used an API or an order-fulfillment company" ignores that, at that time, they very likely were lacking the…

Lots of people make bad business decisions, and I don't want to give them a hard time about the shipping company they chose, especially when they were willing to share the experience in a blog post.

What worries me most is the internal accounting that gets them to $500. I am pretty stingy with the money I spend on my hobby projects, and sometimes it is fun to build things that I could buy for slightly more money. That's ok when the goal is entertainment. If I was at work I would do completely different math, and probably consider a few hours work to save $20 to be a complete failure.