Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Shopify is processing over $260k per minute right now (shopify.com)
333 points by MichaelApproved on Nov 25, 2016 | hide | past | favorite | 108 comments


As someone who has dedicated the last few years of his life building a dev / design shop around Shopify, seeing these kind of numbers makes me very happy.

Looks like we placed our bet on the right horse.

imho xal / Tobi should get far more recognition for what he built. Shopify is really quite remarkable and in a league of its own.

Curious if they'll ever beat Amazon / Alibaba volume.


Amazon sold 426 items per second Christmas 2013. Their average order value in 2012 was around $47. Assuming the average value to be the same, it works out to around $1.2m per minute.


If the average order value is $47, for the average item value to also be $47 it would imply the average order size is one item. That seems highly unlikely.


They're up to $274k/minute. That's $16.5 million/hour. I wonder how Cyber Monday will do compared to Black Friday.


.


We are a public company, no guessing required. Here is the relevant press release and quote:

> Gross Merchandise Volume (“GMV”) for the third quarter was $3.8 billion, an increase of 100% over the third quarter of 2015. GMV processed through Shopify Payments grew to $1.5 billion, which accounted for 39% of GMV processed in the quarter.

https://press.shopify.com/releases/shopify-announces-third-q...


I appreciate that Shopify took a slightly more conservative approach to growth resulting in less feast / famine.

Their RoR stack might have gone out of fashion but it seems like they consistently made the right decisions considering a lot of their customers rely on the Shopify platform for their livelihoods and not just to express their opinions in 140 characters.


Fashion? Fashion is a terrible way to make a technical decision. I'll take a Rails stack over the JS framework(s) of the millisecond any day.


It's not terrible. It is an important factor when considering the pool of devs that know said technology. It's the reason why I've chosen PHP for some projects in spite of my abject hatred for it (for purely aesthetic reasons mind you.)...lots of devs are very familiar with C-like languages.


Except Rails took off with the same fashion and hype and hipsterism. It's just older.


THIS, even rails is not cool anymore. It just fucking works


That is really low, for context a teeny tiny ISO does $1 billion in processing volume a month, so your comparable to one of the hundreds of 2 to 4 employee ISOs out there.


Moyta: Where did you get that $1B per month figure? That seems very very high.

Cheers to the CEO of Shopify answering a question on Hacker News and on Black Friday - that is awesome!


There are 2-4 employee companies processing $1B+ per month? Source please.


No written source to provide, but I know of a 2 person shop that provides merchant services (ISO) processing ~$400MM/month. Not $1B, but is certainly possible I suppose.


forgive my ignorance but in this context what does ISO stand for?


Independent Sales Organization [1], Shopify for example is partnered with Stripe who is an ISO. Neither push big volume so their platform costs are much higher than their competitors.

[1] - https://en.wikipedia.org/wiki/Merchant_account


Really? Stripe doesn't have big volume? That seems... surprising


Actually, I gotta retract from my other comment.

It seems that both Adyen and Brainsomething processed over 50B last year whereas Stripe didn't.

Stripe didn't publish numbers (I'd dare saying because they are very small and that would scare big companies away), but the valuation at 5 Billion dollars indicates that they are less than 1/10th of the aforementioned providers.

Conclusion: Stripe = The hype underdog with strong marketing :D


Not at all. It only depends on your definition of big volume.

Source: By some metrics my company is > 1% of Adyen... or Stripe... or whatever competitor we decide to use at the moment.

If a single company can be as much as 1% of an entire payment provider, they gotta be quite small, right? Well, that's what I think every day when I see news about them :D


nobody buys at the same rate as black friday all year round. doesn't extrapolate to the entire year.


Cyber Monday is a myth. The highest shopping day online is about 2 weeks before Christmas, usually corresponding to ensuring shipping arrives before Christmas.


I work in retail e-commerce in the UK.

Sorry, but you're just outright wrong.

Cyber Monday has set sales records every year for 3-4 years in a row.

Why? It's often been on, or very near, the last payday before Christmas for people on monthly salaries.


At my old company BF always set a new record. Then CM would beat that.

Then almost every day over the next 2 weeks would beat that.


Still worth comparing. And now, I wonder what they'll pull about 2 weeks before Christmas.

I'd be great to see a chart with this data. I wonder if they'll release it or if someone is writing a script to scrape their data.


Uhh what? Do you have any data to back up this claim?


They used to say Black Friday was a myth too but you talk about a thing often enough and sometimes it comes true.


Cyber Monday was(sold it) the highest single day revenue my e-commerce had every year since 2011.


Not to piss on anyone's parade, but for comparison, Alibaba (Tmall and Taobao, via their own Alipay platform) did about 50bn RMB of sales in the first 2.5hrs of this years 11.11 sales festival.

That's roughly US$50m a minute. In the first hour of the promotion they did about 2/3rds of that volume, which is insane.


When I saw the live stream, the first question I asked myself was: I wonder if they're filtering pornographic products and sex toys from that feed?

The answer is no.


Somebody in ops is shitting bricks right now. Heart goes out to you brother/sister.


This is the production engineering pod right now: https://twitter.com/tobi/status/802160212249604096


Do they always work in this room so close together? How can anybody concentrate like this?


0_o Why so many employees? In the context of the volume you do I don't see how you'd turn a profit with that many employees unless you get the tax credits other francophone software groups like Logivision get and more.


For context, shopify has pretty high monthly fees as well as taking a cut of sales. Last I saw they had over 200,000 shops, minimum $30/month (up to $300) means at LEAST $6 mil/month from monthly fees, plus some percentage of the 3.8 billion/quarter gross sales they process... whatever they get to keep after paying strip and the credit card companies. Even a half a percent means $19 million per quarter.

So if you are a startup, growing fast, and bringing in well over $12mil/month... a lot of employees makes sense.


> Total revenue in the third quarter was $99.6 million

https://press.shopify.com/releases/shopify-announces-third-q...


Wow. Internet startups FTW.


Why all the hate on Shopify, from this and your other comment? They seem to be doing a pretty good job. Also, what does the "other francophone software groups" comment mean? Is Shopify francophone and what would that mean?


Not sure if you're joking, but if something happens that takes shopify out just 15 minutes you could pay for that entire room's annual salary.


Are you kidding? 10 minutes @ $260 K / minute would pay for everyone's salary in that room.


Neat D3 based globe component, don't suppose the source is anywhere?


I'd like to know as well. Cool ASCII art in the Console.


I find this page very misleading, only the top-right counters are correct. The products are just randomly choosen from random shops as seen by the function name in the code: `fetchAndDisplayRandomProducts`.

EDIT: I was wrong, see response in the comments


Hey! The products actually are recently sold by merchants who opted in to be featured! We populate the first 5 and then new ones come down through the websocket connection


I'm a Shopify store owner, how do I opt-in to be featured?


If you log in to Shopify you'll notice a card on your admin home that says "Reach a wider audience during Black Friday and Cyber Monday". That card will give you more information and allow you to opt in.


mea culpa, I just skimmed through the code because I think it would be interesting to graph that stuff.


You guys should make that clear, wouldn't want that information to spread.

This seems like a great start up to learn from, I wonder how they carved out an ecosystem from under amazon and such. Seems awesome.


Shopify is no longer a startup, they're a public traded company on the NYSE: https://www.google.com/finance?cid=990038606551779


Being publicly traded doesn't automatically mean you're not a startup. Startup = growth, and there are plenty of other things aside from stock price or employee count to look at. For example, Tesla was public when they just made a single car model, but now Tesla is an energy company and expecting rapid, multi-year exponential growth from new markets. That kind of search for product market fit and growth from a company is much more of a startup attribute than anything else, and looking at how a company plans to grow is a much better indicator of how we should use 'startup' than "is it public or private".

http://www.paulgraham.com/growth.html


To be fair, it wouldn't be the first time somebody took an existing function on a page and repurposed it to do something other than its original name.


The two hard problems in computer science strike again ;)


Woa! I wonder what this chart would look like for Amazon


Amazon used to show real time sales stats on their homepage during the holiday season. I think it was the # of orders, IIRC?

I wish they still did that, it was mind boggling even many years ago.


higher, but not much much higher. A lot of small things combine into something quite big!


Amazon sales in 2015 were $107.01bn, so let's say they sell 1bn on Black Friday. This would be around 300 times the numbers shown.


I don't understand how you're calculating this. I think you missed a couple of zeroes?

If you take that $260k/minute number and assume it's true for the whole day (clearly it won't be overnight, but bear with me), that's (260,000 * 60 * 24 =) $374M. A third of a billion, a far cry from 1/300.


You are making your calculations based on the peak minute of the day, I'm checking now and it's around half that and it's middle of the day in the US. I imagine during the night is several times less than 260k.


Of course, but not 100x less. Amazon's number is definitely substantially higher, don't get me wrong, but 300x is just not right.


I underestimated Amazon sales on my initial comment, it's around 3bn: http://fortune.com/2016/11/25/black-friday-online-sales/

The numbers are big but far from giants like Amazon and Alibaba: http://www.cnbc.com/2016/11/11/singles-day-news-alibaba-pois...


I don't understand why this is being downvoted. Amazon is a big store but Shopify is "basically" the platform for every other ecommerce store, so is every other online store combined. It makes sense that it'd be close to the volume of Amazon.


lol, not true.


Is this actual realtime data or estimations updated once in a while and incremented on the client?

Either way, looks neat (and some crazy numbers)


Actually real time streaming over a web socket.


shopify is still written in rails? :)


Absolutely!


this is really nice. For all the news that you hear about ruby and rails, that is slow, that is bad, that everyone leave the language. I found it really nice. I do follow and see all the video conf regarding ruby, and I do hope that next year, jruby will be 3 times faster, and jruby+truffle will be able to run normal ruby gems. As for ruby 3, to be 3 times faster, for that we wait and hope


I'm sorry to be this guy, but what a simple checkout store that Shopify is (nothing wrong with that) has to do with ability to scale and add more servers?

Whether this was written in PHP, .NET or Ruby.


What an arrogant comment, "simple checkout store", yes if they where doing it for one customer perhaps but they are handling inventory, sales, displaying products, theming, delivery, invoicing and payment processing for thousands of clients and millions of customers...

"simple checkout store" - brilliant.


Ha! Shard on shopId. ECommerce can be hard, but there's nothing special about what you descibe!


I'm going to attempt to bring this back to a positive and constructive conversation.

From the sounds of it we are all devs here. And, as experienced devs, we all know that saying "you just have to ..." will be your famous last words.

Adding more servers, as many of you will know, only works up to a point. From there you have to further build out databases, redis, memcache, load balancers, etc. As much as we like to fantasize about a "go faster" button which just adds more compute, often the amount of memory or CPU you have isn't the bottleneck.

When it comes to sharding, you're not wrong. Sharding off of a primary key is something that is easy to grasp. Hell, it's not overly special (especially these days). However, that doesn't mean it's easy.

If you think building a commerce platform is easy, I'd encourage you to apply. Maybe you have the type of skills we are looking for?


But looks like there are some bits of Go in there somewhere... https://github.com/tobi?utf8=%E2%9C%93&tab=repositories&q=go...


Yep we run Go for some infrastructure and ops kind of things, but the main app you hit when you visit any Shopify store is 100% Rails


Amazing. They need some outreach in Africa and South America.


Does this include POS sales?


it does. All Shopify sales channels are included. Web, Pos, Native Apps SDK users, Facebook, Pinterest, Twitter, Houzz, etc


No clue, hopefully xal will comment on this seeing as he works for them.


He's Tobi Lutke, founder/CEO/Coder.


"Rails doesnt scale" mimimi


Are they using actioncable for this?


No, ActionCable for this one, it's a Go service that's handling the websockets.


Do you know any reference on how to implement something like this with Rails (go + rails)?


Did you guys talk to your legal department before going live with this? As Shopify is a public company, showing sales/revenue stats like this in real time might be in violation of SEC regulation FD if you did not alert all your investors.

EDIT: Actually, since they are a Canadian company they are exempt [0] from regulation FD. Interesting!

[0] - https://www.sec.gov/Archives/edgar/data/1594805/000119312515...


So it turns out the SEC lets you use even things like Twitter to give out information, as long as it can go to every shareholder at the same time [I AM NOT YOUR LAWYER, IGNORE ME].

see: https://www.sec.gov/News/PressRelease/Detail/PressRelease/13...


You still have to formally tell your investors that you will be using Twitter (or some other internet platform) to disseminate said information.

This Shopify data could definitely be used to gain an edge in trading their stock. You basically compute projected earnings based on the sales per-second rate reported on the site and compare against their earnings guidance from the previous quarter (or analyst predictions). If the projected earnings go below the projection, you sell, otherwise you buy.


This is on their website. Therefore it's already an accepted source of information for stock related material. Since it is public and went out to everyone at the same time, it should be fine.


Looks like it’s been up for at least a year[1], so I expect they’ve thought about that.

[1] https://news.ycombinator.com/item?id=10637715


You'd think it'd perform better, on my i5 with 16GB of ram its laggy as can be.

Edit: Keep the downvotes up, but having this site eat a full core is a bit ridiculous. Similar sites are much more performant.


Any examples of similar sites? It smashes a core of my laptop (i7 / 16GB) as well, but every site I've been to that looks like this (3D realtime) does the same thing, so I'm used to it.


Works smoothly enough on a MacBook Air 1.6GHz with 4GB RAM and Windows 10.


Works smoothly for me - i5 and 16 gigs of RAM.


Doesn't for me. i5 and 24 gigs of RAM.


see, you can still scale and make money with Rails in 2016, even if it's not cool anymore :)


Additionally, stores that run on Magento and WooCommerce would agree that the coolness of the language (PHP) has little to do with making money. :)


They tried JavaScript too and went back :)


Do you have details on that?



for context, sorry if ignorant question but what else do they write with when full scale, I thought twittter was written on rails as well


I think it's fair to say that there are high scale websites built using any stack, mostly using more than one technology.

I was just being ironic in my previous comment.


Twitter moved a lot of their codebase from RoR to Scala a few years back.


UX orgasm, very powerful with very few elements


Hate to rain on this (TBH I don't):

Shopify has raised (per Crunchbase) about $122M in funding.

At the current rate, they will see $374M in GMV on their best day of the year.

Shopify has between 1K and 5K employees (again Crunchbase). Let's make the most conservative analysis and assume 1K. If they are paid $20 per hour (low estimate I think), that is $41M in employment costs per year. With taxes, office rent, healthcare, etc, lets just lowball again and call it $50M/yr for all the people costs. Again I think this is a huge underestimate if the crunchbase number is accurate.

Lets assume no marketing, hosting or other costs of any kind.

The internet claims 100k stores use shopify. Lets assume they all pay for the standard tier plan, and pay month to month. That is $79 * 100k = 7.9M per month in revenue. Again, huge over estimate, since I would guess 95% of the customers use the cheapest tier, and the largest companies most likely negotiate better than advertised fees.

Finally lets assume their BEST day is only about twice as good as their average day. This is probably overestimating their revenue by about 10x, at least.

That would mean they are bringing in well under $47.4M in gross revenue, on over $50M in fixed costs.

So in other words, even if black friday were half the days in the year, they still would lose substantial money. This is after 12 years of business.

It's unlikely they will ever turn a profit. If they were going to experience some kind of exponential growth to defeat this analysis, 12 years of the fastest growth in the internet have already gone by, so it seems unlikely going forward.


Shopify is a public company. You don't need to speculate on their revenue or expenses, you can just look it up. They earned $99.6 million in Q3. They expect revenues in the range of $379 million to $381 million for 2016. And they are still losing money, although getting close to break even. Their market cap is 3.86B, so 10 times revenue, so the market is expecting them to continue to grow rapidly.

https://investors.shopify.com/events/Events-Presentations/de...


Hey, you're right!

https://www.google.com/finance?q=NYSE%3ASHOP&fstype=ii&ei=IE...

Revenue growth with corresponding increases in costs and expenses. NOT getting closer to profitability, in fact operating losses are increasing in proportion to revenue (almost a linear relationship in every quarter and year in the reports).

Again, this is a failing business. 12 years in and no profit, profit not within reach, costs growing in lockstep with revenues.

EDIT:

So I was curious how they had such high revenues when clearly subscription fees can't account for it. Found this in the financial statement:

"We principally generate merchant solutions revenues from payment processing fees from Shopify Payments. Shopify Payments is a fully integrated payment processing service that allows our merchants to accept and process payment cards online and offline. As a result of the launch of Shopify Payments in August 2013, we have seen significant growth in the revenues generated from our merchant solutions. In addition to payment processing fees from Shopify Payments, we also generate merchant solutions revenue from transaction fees, Shopify Shipping, Shopify Capital, referral fees from partners, and sales of point-of-sale ("POS") hardware. Our merchant solutions revenues are directionally correlated with the level of GMV that our merchants process through our platform. Merchant solutions revenues increased from $34.5 million in the six months ended June 30, 2015 , to $77.0 million in the six months ended June 30, 2016 , representing an increase of 123.4%"

So they are counting CC processing fees towards revenues. I suspect that CC payment processing fees shopify pays are approximately equal to what they are charging, so this is just a revenue game where they charge 3% to customers, pay almost 3% to the CC companies, and book revenue that can never be profitable.

EDIT2: So found the cashflow in the statement. They charged $76M for these fees, and paid out $56M. Not bad actually! However, that $20M in net is still much lower than the costs to generate the business, $56M just in marketing.


> So they are counting CC processing fees towards revenues. I suspect that CC payment processing fees shopify pays are approximately equal to what they are charging, so this is just a revenue game where they charge 3% to customers, pay almost 3% to the CC companies, and book revenue that can never be profitable.

You don't understand the meaning of the word revenue.


Maybe, I think I know what it means but I'm not an accountant.

I know that revenue is often quoted as some kind of measure of the success of a business, when in reality it's extremely easy to run up revenues if you are running at a loss. In the most trivial case, you can pay $1.10 for $1 bills, for example.

Ok, so shopify isn't doing that, but it's hardly different, just a more complex situation. They pay $50 for advertising to get $75 in revenue, which they immediately pay $55 for. That's $105 paid out to buy $75 in revenue.

It would certainly be cheaper to just pay $1.10 for each $1 in revenue. I'm sure the folks at shopify don't think of it this way, and I'm sure they genuinely believe that they will someday change the economics of this and become profitable. However, the balance sheet doesn't support that belief, for me at least, and it sure looks like every customer they maintain leads to more money they are losing.

Perhaps they can become profitable with enough economies of scale. Again, I would question how much larger they would have to get to hit the inflection point, which they haven't done, at least in the financials reported, so far losses are scaling up linearly with revenues (which is to say operating margins are a steady negative % against increasing revenues).

Since they have been around for a long time (12 years) and those same 12 years have seen an absolutely huge increase in ecommerce, it seems like they have failed to hit the required scale despite the most favorable conditions you might hope for. That doesn't make me very optimistic.


Assuming you based this off their financial statements, then everything you are saying is pretty much 100% on the ball with regard to interpreting their cash flow.

> That doesn't make me very optimistic.

I would say that the market is "on your side" at least for now, as people on here have been starting to compare them to Amazon which is, in terms of market cap, 100x higher valued than Shopify and really not a fair comparison. However, let's not forget that the reason Shopify is a 3.5bil company (and not worth say "only" 500mil) is because they have eyeballs on them right now, and quite a few at that. The market isn't sure how they plan to use those eyeballs, but there's enough visibility to give them a chance at being a strong market player in the future. Also yes, despite being publicly traded, most investors would consider them still to be startup based on where they are at with their business model.


I think they're required to do it that way if they're preparing their financial statements in accordance with IFRS or U.S. GAAP.

In the case of processing fees, I'd expect those to be what makes up most of the 'Cost of Revenue' line on the statement. In a company selling physical goods, it would be called 'Cost of Good Sold'. Basically, the cost of the raw materials you had to acquire to build your products. Because of this, they tend to scale in proportion with revenue, and it looks like this is the case for Shopify. Their cost of revenue is a little under 50% of revenue, so I don't think it's fair to say their revenue can never be profitable.

In terms of profitability, their net loss isn't that big as a percentage of revenue. They could probably, if they wanted to, let revenue grow for a while without increasing staffing levels or advertising spend too much - it wouldn't take long for them to be profitable. But if you've got enough cash on hand to run a relatively small loss in order to continue growing, solidify your position in the market, and expand into new markets (like they've been doing recently with Shopify POS) it makes sense to do so. They have

Right now, they have $409.39 in cash and short term investments. That money isn't going to help you much just sitting in the bank, and using it to grow the company will probably result in a significantly higher ROI than other things they could invest the money in. The way it looks to me, they're not running a loss because they have to. It looks like they've made a conscious decision to do so, and it looks like that decision is a good one based on the company's financial position.




Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: