Contents Preface Peecho Introduction About the company What is Peecho? Mission Business model Brief history Problem solved Dream Team Developer community What others say Tweets Customer stories In the press The inside story Beginning Business model Execution Community The investor game, part I Technology Expensive Stuff Does Not Scale If You Are Slow, You Can't Grow Sure, But What Does The Architecture Look Like?
4 5 6 6 6 6 6 7 7 7 10 10 12 12 13 14 16 16 18 20 22 24 26 26 26 27
Preface The guys from Peecho and I go back a long way. They were part of our incredible team at Albumprinter, the successful photo book company that was recently acquired by Vistaprint. I recall that the penguin logo showed up everywhere in the building, from meeting rooms to even the whiteboard in my office - years before they used it for their start-up. When they started with Peecho, I was very excited about their solution. In our Albumprinter days, we could service only a few of the largest companies, while Peecho can serve anybody - big or small. This excitement was the reason why I invested in Peecho and agreed to join their advisory board.
apps and games to hundreds of production facilities all around the world - in the simplest way on the market today. The possibilities for customized merchandise for content - be it art, photos, documents, games, TV shows or events - are endless. Digital content exchange is exploding, creating a new market for on demand production of 2D or 3D printed objects. This market needs a new, simple, global infrastructure. That’s why I put my money on Peecho.
Joris Keijzer Peecho transforms pixels into print. Their platform is capable of connecting thousands of websites,
4
Former CEO of Albumprinter
Introduction This publication is a classic example of scope creep. First, we needed a fake publication for a single print proof for every bookish product in our catalogue. So, we needed lots of images with colors and quite a few pages. Then, we figured we could use these proofs to compare the quality production facilities. It started to dawn on us that it must resemble a real publication, instead of a colourfull mess. Then, we found out that we also needed a demo publication on our website, exposing it to all our visitors. The stakes just got higher. It had to look good on a screen.
Then, our sales people wanted demo products that they could show to potential customers. This means that the printed versions really needed to reflect our quality standards. Then, the commercial guys wanted the publication to be a sales brochure that explains who we are and what we do. Now, the texts had to make sense, thus raising the bar even higher. The grand total of all of the above is what you are holding in your hands right now. It is a miracle.
5
About the company What is Peecho? Peecho is a free service that allows anyone to sell professionally printed products from their application or website, and make a profit with every sale. Customers simply sign up, get the print button code and embed it in their website. Visitors can then buy digital documents or images as photo books, reports, magazines, canvas prints, and more. An API is also available.
Mission Peecho’s mission is to allow for the transformation of digital information into other forms. We want virtual information to be readily accessible as tangible objects - but only by means of local production, to minimize the environmental footprint.
The cloud print button is powered by Peecho’s global cloud print network that operates as the intermediary between many applications and only the best production facilities in the world. It facilitates ordering, conversion of user-generated data to print-ready files, production and shipping. Using cloud technology, production orders are routed to suppliers that are located as as close to the recipient as possible,minimizing transport overhead and ecological impact.
Business model We are a B2B service. We charge wholesale prices for every product type. These wholesale prices include our profit margin. Our customers can decide on their own retail price - that’s what a consumer pays for a product. The difference between wholesale and retail price is their profit.
There is no minimum volume or investment required, eliminating your risk completely.
6
The cloud print button is used by hundreds of websites and apps. We collect consumer payments for the orders and compensate the customer with their profit margin afterwards. Next to that, selected API
users are charged monthly for the orders they placed in a pay-for-use model. Brief history Peecho was founded in November 2009 by Martijn Groot and Sander Nagtegaal. Six months later, the company got a lot of attention after the rapid launch of the cloud print API and won the award for Best Business Model at The Next Web conference in Amsterdam. The rest of the year was used to perfect the infrastructure with a few lighthouse customers. Then, it was time to scale out. By the end of April 2011, we launched an embarrassingly simple product on top of our platform. It was an embeddable cloud print button. We pitched it as “the Facebook Like for cloud printing”, which was picked up by large technology blogs.The response was quite overwhelming. Our customer base increased tenfold in one week. Peecho was recognized as most innovative e-commerce company in The Netherlands by the jury of the Accenture Innovation Awards in October 2010. A month later, Peecho closed an investment round with Peak Capital and DHG, allowing the company primarily to increase their sales force in the year to come.
Problem solved Digital data is created on a massive scale: tablet publications, game avatars, photos, blog posts, tweets and so on. However, millions of images, documents and other digital publications are still waiting to be sold as tangible products. Why don’t content owners sell their digital assets as physical objects more often? The answer is simple. The execution is highly complicated. As a content owner, you would have to: • find yourself a production facility, • discuss pricing, • buy software licenses, • create an infrastructure, • integrate your app, • translate crappy digital data - that was never meant to be printed into ISO certified print-ready files, • and transport the results into the existing facility workflow. You would then have to closely monitor production and distribution of every order and offer customer service. The risks are huge: to get a reasonable price from your print facility, you would be faced with an up-front demand for a minimum required number of sold products. Internationalization, switching suppliers or extending your product portfolio can mean that the whole ordeal will start over.
money. We released our cloud print button, a free service that allows any content owner to start monetizing their content. Dream Our dream is bigger than that. We are going to disrupt the entire publishing market. Historically, publishers started with a vaporware, printed publication and subsequently surrounded it with digital services. This multi-billion dollar, heavily fragmented market is currently collapsing. Tomorrow’s publishing world is all about digital assets published on tablets, e-readers, mobile devices, websites and games. As a result, a smaller market will emerge for high-quality, environmentally friendly, on-demand print services in 2D and 3D. An entirely different infrastructure is needed to support this market. Peecho will be that infrastructure.
So, we decided to solve this problem by helping other businesses to make
7
Team Martijn Groot Martijn Groot is CEO of Peecho and is known for his successful online and offline experience in commercial sales and startups.
Sander Nagtegaal
Prior to founding Peecho, Martijn was Commercial Director and led all B2C marketing and sales for AlbumPrinter, the first European company to enable consumers to create photo books using their own digital images. The company grew by over 22,000% during its first five years.
Prior to founding Peecho, Sander was Chief Architect for AlbumPrinter where he was responsible for IT vision, strategy, innovation and business development of everything from consumer software to plant automation. He led a team of 30 developers and architects for over four years.
Before AlbumPrinter, he spent nearly seven years working within The Kinnevik Group in a variety of roles that include Manager of Partner Sales for Tele2, Marketing Director for Alpha Telecom and Sales and Marketing Manager for Tele2. In that last role he helped prepare and launch their first MVNO in The Netherlands.
Previously, he founded and led Centrical, an international consultancy firm focused on IT architectures, technical management and strategy that counted Akzo Nobel, UMC, Hogeschool Utrecht, HKZ, AlbumPrinter, Evident Interactive and others among its clients.
Martijn holds a bachelor’s degree in Marketing from the Avans Hogeschool, the international business school located in Breda, The Netherlands. More importantly, Martijn loves to spend time with his daughter and her mom. He plays squash, is part of an indoor soccer team and goes kitesurfing or snowboarding as often as possible.
Developer community Our developer community executes the heavy lifiting at Peecho. The following guys were part of Peecho from day one.
10
Sander Nagtegaal brings a history of successful experience in IT leadership, printing and startups to his role as CTO of Peecho.
Sander holds a Masters of Science degree in geophysics from the University of Utrecht. In his spare time, he enjoys mixed martial arts (MMA) classes and teaches kung fu in Amsterdam.
Marcel Panse Chief architect and all-round cloud computing guru.
Willem Vermeer Java sage, master of ceremony and our data issue savior.
Stephan van Eijkelenburg App authority and chaos monkey.
Henrich Volkshofen Our German print expert with over 65 years of experience.
11
What others say Tweets
The Next Web @thenextweb “This startup will rock your world #peecho”
Wimer @monokai “Found the PDF-to-print web service I was looking for: Peecho!”
Werner Vogels @werner “If You Are Slow, You Can’t Grow - Peecho Architecture - scalability on a shoestring”
Michael Josefowicz @ toughloveforx “Peecho receives Accenture Innovation Award for best e-commerce concept!“
Jeff Barr @jeffbarr “A very informative article - “Minimizing Downtime on Amazon AWS” by @peecho”
Steffen Konrath @ stkonrath “Peecho transforms pixels into print: license to print nets it $750,000 in funding”
Manon van Essen @ manonvessen “4 secrets of global domination and 10 slides to realize why you like Peecho so much!”
Jons Janssens @jonsjanssens “Congratulations to @ peecho for best business model!”
Herbert Pesch @herbertpesch “Peecho shows the math for start-ups @ rockstart”
Mathys van Abbe @mathys “w00t go dutchies! RT @ TechCrunch: Peecho Lands $750,000 For Its Cloud Print Button“
Peecho @peecho “Today, we received a great gift from Twitter: from now on, our username is @peecho”
12
Customer stories Issuu Issuu is the leading digital publishing platform delivering exceptional reading experiences from magazines, catalogs, and newspapers. Millions of people have uploaded their best publications to create beautiful digital editions.
Hyves Hyves is the largest local social network in The Netherlands, boasting over nine million users. They have been holding off global competitors like Facebook for years. Their social game platform and payments structure have been exceptionally successful. Next to that, Hyves is known for their corporate responsibility. A recent initiative is the release of an educational publication aiming at primary schools. In ten courses, children are explained how to use the social network safely.
Issuu is not only the fastest growing digital publishing platform in the world, but also a very popular destination site where people are engaging with the web’s best publications and where many publishers build their audience. One of Issuu’s latest additions is the print feature for publishers. Now, every digital publisher has the opportunity to transform their digital publication into a real, tangible product. Options include magazines, glossy paperbacks and hardcover books - all printed on demand, regardless of size or number of pages. Issuu’s print partner is Peecho. Their cloud print button connects digital publishers like Issuu with a large network of the world’s best production facilities. By analyzing the specifications of any of Issuu’s documents in real-time, the cloud print button automatically calculates the accurate, lowest price for a printed version and shows the result upon display.
While the company’s gesture already resulted in quite some enthusiasm in the press, it gets better. Although Hyves primarily offers the publication as a PDF file for download, it decided that a printed edition was necessary for at least part of the audience. The simplest way of doing that is adding the Peecho cloud print button and so they did. Hyves really understands what corporate responsibility means. They do not waste paper, but offer their educational publication for print on demand only. Boris van der Ham, Dutch politician for D66, called it “a great initiative”.
Mik Stroyberg, Director of Consumer Engagement at Issuu, says: “Due to the diverse nature of Issuu’s publications, transforming our pixels into high quality print is really complicated. Peecho does a great job by offering a service that used to be impossible. We are looking forward to generating a lot of revenue with their help”.
13
Kodak Gallery KODAK is well-known as an established brand in the printing industry. One of its successful spin-offs is KODAK Gallery, a highly popular sharing and storage medium of consumer photos. More than 75 million people worldwide manage, share and create photo gifts online at KODAK Gallery. Recently, the company started using Peecho to power their new generation of applications.
Today’s world is highly distributed: social networks and mobile devices contain many parts of a person’s life, too. So, KODAK Gallery decided to extend its presence to new audiences.
14
The company started building a series of applications that allow you to create great personalized products right from where your photos are - from within social networks, mobile devices and of course the KODAK Gallery websites.
In the press The following is an article that wellknown technology blog VentureBeat published after the news of the Peecho investment round led by Peak Capital and DHG Holding. Other articles can be found at TechCrunch, TheNextWeb, The Atlantic and many others.
As a result, KODAK Gallery will help you to make it easier and more affordable to capture your best moments in life, from wherever you keep those. You don’t have to bring your memories to the software - KODAK Gallery will simply bring the apps to where your photos are. An example is their iPhone application for sending physical postcards using your own photos. KODAK released it in six different languages, specifically aiming at the European market.
Peecho’s “License to Print” nets it $750,000 in funding VentureBeat, Ciara Byrne, January 21, 2012
Mario Grunitz, Senior Manager Online Sales & Marketing EMEA for KODAK Gallery, says: “The partnership with Peecho allows KODAK Gallery to rapidly develop new apps and distribute the resulting production orders to multiple print facilities all over the world. Eventually, this will result in much lower transportation costs and a significantly smaller ecological footprint.”
Peecho, which just raised $750,000, lets visitors to a website or application transform pixels into print by clicking on its embedded print button. The print button connects to a cloud of print facilities all over to world which can produce any chosen magazine, photo album, poster or book.
For a writer, nothing beats the romance of seeing your work in high-quality print. No screen can compete with the silky feel of the paper, an excess of glossy photographs and that special smell of a new page.
Founders Martijn Groot and Sander Nagtegaal met at photo book printers AlbumPrinter which ran huge printing facilities churning out 16,000 photo books a day, each one different. “We saw that
people were telling their own story in digital media,” says Groot, “but still wanted physical products.” Most online content is not print ready. “PDF doesn’t have a spine,” says Groot. There is a bewildering range of digital publishing file formats. This is the reason that there are plenty of sites that let you print a photo album or maybe order a magazine on demand, but most services only print one type of product via a single print facility. Peecho connects to a cloud print network of specialised printing facilities all over the world and aggregates orders from different customers. Most print facilities currently have excess capacity and Peecho can connect a new facility within 2 weeks, a process that would formerly would have taken months. The company takes a markup of each order on top of the wholesale printing price and also provides a white-label solution. Most of the sites and applications which currently use the print button have never had print products as part of their offerings. Peecho will launch a service in February with digital publishing platfom Issuu, which has 50 million readers and adds 201,000 new titles every month. A visitor will be able to choose to print a magazine, paperback or hardback in color or black and white.
Only about 5 percent of Issuu’s content is currently available in print. “A lot of titles are never published in print because the volume isn’t large enough,” says Groot. That makes Peecho the long tail of print publishing, or as Groot calls it, “professional printing for the masses.” On-demand printing is still more expensive than off-the-shelf. A supermarket magazine which costs $5 might cost $7 dollars to print on demand but Groot expects prices to drop as volume increases. While Peecho’s customers are still mainly digital publishers, social media services like Walnuts, which creates books based on Facebook content, are increasingly offering print products. Groot claims that while there are competitors in different sectors, for example Fotomoto to print photographs for professional photographers, no other company covers multiple print formats and facilities. Peecho’s new funding comes from Peak Capital and DHG Holding, B.V. and will mainly be spent on expanding the company’s global sales force and scaling up the business. The company is based in Amsterdam, has 4 employees and was founded in 2010.
15
The inside story The following snippets were previously published on Sander’s blog. Beginning Running a start-up doesn’t sound great. I couldn’t tell you what will happen in the next couple of months, the hours are ridiculous and the salary sucks. Still, I love it. Let me tell you about Peecho, so you might see why. The company is run by Martijn and me. Martijn is a tall blond guy, fast-talking and typically very nice to most people that he knows - and he knows everybody. I guess that is what commercial people do. It gets worse, though: unlike me, he is also the kind of guy that celebrates even the smallest victories with the team, sends supportive postcards and text messages at the right times, and inevitably calls you at those occasions of which a subtle protocol requires personal contact. Anyway, I am smaller and arguably less friendly, although I have my moments. The contrast is most striking at sports. Martijn plays soccer in a team of friends, I practice mixed martial arts and kung fu. Peecho does not go back a long way. Martijn and I have not even known each other for very long. We met only a few years ago at Albumprinter, where I was chief architect and he was commercial director. We didn’t
16
speak with each other often. I don’t know why. Techies don’t necessarily get along with people with “commercial” in their job title, but I am not really like that. I get along with many people. It was just coincidence, I guess. Albumprinter was an instant success. To resellers, the company offers consumer software to let anybody create photo books. This includes everything from creating the book to production and shipping from a dedicated factory, a fascinating building near The Hague. It’s like Willy Wonka’s factory, but without the chocolate. Some of you might argue that the chocolate is essential here, but I really don’t think so. What counts is the machinery, the noise, the smell and the magic when a product emerges out of the turmoil. I love it. However, I am not really a print guy. Years ago, I graduated as a geophysicist. These are the kind of scientists that research the earth first and then end up working in a completely different business instead. My different business is the internet. I am a software architect. My own consulting firm, Centrical, was doing pretty well, but I longed to do something substantial for a ompany with a cool product. A product that real people would use and like. Joris Keijzer, the Albumprinter CEO, managed to lure
me into an assignment as chief architect to do just that. His new company was growing at the speed of light, creating a huge technical mess at about the same velocity. He thought I could help - and so I did. For years, I fiercely tried to innovate and keep everything running at the same time. Mostly, it worked. I had a great time. Martijn came in somewhat later, with a specific task. Next to the reseller business, Albumprinter used its own name to test new software directly with consumers. To our great surprise, the brand became rather successful. Now, there was a problem. To our business customers, he company was supplier and competitor at the same time. So, it was decided to launch a new consumer brand instead. It was named Albelli. To launch a new brand like that, you need Martijn. In the years after that, Albumprinter’s revenue skyrocketed. We won every prize in the galaxy for our unearthly growth. You can Google it if you don’t believe it. We were doing well. Extremely well. So, if you have seen a lot of movies, you could have seen this one coming: in every scary movie, there is an opening shot that shows an unlikely happy family. You just know it will all go wrong. The same thing happened to Albumprinter - except for the serial killer.
Although it sounds weird, the problem was the growth. If you produce physical stuff, you need supplies first. Especially around Christmas - the peak season - you really need a lot of paper, ink, people and machines. So, we had to borrow lots of money to buy stuff before we could even start selling books. Then, the economic crisis hit the world. Banks got scared. The result: despite the unprecedented revenue growth, Albumprinter had to cut costs dramatically and consolidate, at least for a while. With expansion and innovation out of the window, both Martijn and I were nothing but a vastly overpaid pain-in-the-ass to the company. It made sense for us to leave. So we did. It wasn’t easy, though. We put a lot of time and energy into the company and many of our friends still work there. Therefore, we wish Albumprinter all the best. When the smoke cleared, we found ourselves on the roof terrace of the public library. You should go there - it boasts a great view over Amsterdam. We were drinking coffee in the sun, discussing new opportunities. Martijn was just about to become the Big Kahuna at a well-known electronics empire, while I prepared to move to San Francisco - the Valhalla of the internet industry. Then, quite naturally, something beautiful happened. While sipping our coffee, a new idea formed. We decided that we were not done with the printing business. In fact, we were just beginning. On top of that, we realized that we would make excellent partners - our respective traits were practically complementary. Right there, on the roof of the library, Peecho was born. I can’t tell you what will happen in the next couple of months, the hours are ridiculous and the salary sucks. Still, I love it.
17
Business model The first weeks of our start-up adventure were rather chaotic. The most complicated thing was to decide what we were going to do. Basically, we needed to get our story straight. Our hectic brainstorm sessions lead to some outrageous business proposals that would never see the light of day - and never will, I hope. Initially, we thought it was best to keep the results of those sessions a Big Secret. The master plan was very simple at that point. We would secretly invent a killer product and then rule the world. Luckily, we changed our minds about the secrecy. Stealth mode doesn’t work. It’s all in the execution - what will ultimately make the difference between success and failure isn’t the idea itself, but the ability to execute and dominate the market very fast. We stopped worrying about somebody stealing our ideas and decided to publish everything that we were doing. It worked. Many smart people approached us, triggered by the stuff we wrote on the web. They helped us figuring out what to do. However, even with lots of help, it’s incredibly difficult to get a story straight. It takes forever and you never really get there. When you think you do, some show-off then comments on your business plan and you will have to rewrite it. Again. And again. But without all this help, our plans would never have converged to something useful. While frantically rewriting our business plan, we realized that making money is actually quite hard. It didn’t used to be like that, though. The idea behind a successful start-up was once simple. 1. Build a really cool service. 2. Offer it for free. 3. Get yourself a whole lot of registered users. 4. After a year or so, start bombarding them with expensive advertising. 5. Buy yourself a big yacht.
18
Yes, those were the days. Then, a couple of bubbles burst and along came a crisis. Advertisers cut back their spending. All there is left now are paid subscriptions and selling stuff. Our time at Albumprinter showed us that selling stuff is good, but selling personalized stuff is even better. So, that’s what we were going to do. We started off by thinking that we would offer a couple of mobile applications to sell personalized photo products. However, after careful examination, the execution promised to be complicated. It is not easy to start selling personalized, physical goods. It requires a large up-front investment. It begins really simple, though. First, you need to choose a relevant product, like a glossy magazine, greeting card or a t-shirt. Next up is creating a user interface. Then, the real trouble begins. You need to find yourself a print facility, discuss pricing, think over your infrastructure, buy the relevant software licenses, worry about product definition formats, figure out how to offer themed templates, choose between translation to either PDF or PPML and transport the resulting files into the facility workflow. After that, you cross your fingers, hope that all goes well in production and wait for status updates about shipping and such - if those come at all. The worst part: to get a reasonable price from your print facility, you will be faced with an up-front demand for a minimum required number of sold products. At that point, you are not even done. Let’s assume all goes well and it’s time to extend your product portfolio. You call your print facility and ask them if they can print your new product of choice. Unfortunately, the answer is no. So, you get yourself a new factory somewhere. Tough luck - the new print facility doesn’t comply with your existing infrastructure that you put so much effort in. Now, the whole ordeal starts over. Ouch - I loved Ghostbusters, but Groundhog Day was a nightmare.
In short: our ideas required lots of infrastructure and at least a couple of print facilities. That’s a lot of work for just a few mobile apps. Maybe somebody else could do the work for us. Apparently, there are software providers that can help - but their licenses come at a huge price. That investment was unacceptable As an alternative, we triedto find somebody who could offer this kind of infrastructure as a service, preferably pay-as-you-go. We failed miserably. There was nothing to be found. Boy, were we lucky. That night, Martijn and I celebrated our failure. We got pretty drunk. Sometimes, it’s awesome just to fail. In this case, it meant we had found ourselves a business model. After all, to make money, you will have to search for the things that are hard to do – that’s where the revenue is. So, instead of focusing on our initial concepts, we quickly decided to start scratching our new itch. We were not just going to build applications and sell stuff. We were going to help other businesses to make money instead. We were going to offer print as a service, without a minimum required volume, pay-as-you-go, no investments needed. Our story was straight - for the time being.
19
Execution After we figured out what we were going to do, it was time to move on to the execution. That meant paperwork - lots of it. The complicated ritual of establishing an official company was the worst part. We spent heaps of money on lots of paper with only symbolic value, written by a bunch of lawyers, a notary and the chamber of commerce. Before we knew what happened, we were proud owners of no less than four companies: two personal holdings, an investment organization and the actual company. Isn’t that weird? You ask for one, but you get four. The gruesome administration effort was soon forgotten, when the rest of the work doomed on us. Starting a company means a lot of do-ityourself. Next to marketing and sales, Martijn also deals with the administration, drives the car and cooks a fine dinner, too. I am the architect, evangelist, fitness instructor and graphic designer at the same time. Together, we make one really well-rounded employee, but four hands type faster than two. It’s more fun this way, too. Next to a master plan and skills, every company needs a symbol. We have our penguin. I designed the penguin during my graffiti years. As my little protest against authority, I painted it everywhere. Eventually, I transformed into a law-abiding citizen, and I
20
started to draw the penguin on work-related whiteboards instead. It wouldn’t surprise me if you could still find a few in the Albumprinter building somewhere. So, it’s a dodgy underground-figure-turns-commercial story. I apologize and promise not to make a movie. I may write a book, though.
In digital form, the penguin turned out really basic, because we read the book “Getting Real” and lacked money at the same time. Don’t confuse the title of that book with the Ali G election slogan - I did, once, and got aughed at by my girlfriend. Anyway, among other things, the book claims that restrictions are good - run on limited resources and you will be forced to be more creative. So, I never use complicated, expensive programs for graphic design. I know that it becomes a mess if I do. To avoid brushes, filters and fades, I constrain myself to a cheap technical diagramming application instead. It’s very simple. It doesn’t support any funky stuff. We use it for all our graphic design now.
When it was used in vandalism, the penguin had no name. According to my marketing-savvy partner Martijn, a brand that is merely visual is hard to communicate. I can see his point. After all, he is the guy who suffers the most - when he isn’t doing the paper stuff, he is in phone calls all day. Therefore, the search for a really cool domain name started. We didn’t think of a name first. We approached it the other way around. We checked the lists of recently deleted dot-com domains and found our name quickly. “Peecho” is short, snappy and it sounds like it could be a penguin’s name. Then, we checked Google to see if it meant anything. After all, you don’t want to end up with a company named after a criminal organization or a fetish sex toy. No dubious contamination was found, but we quickly noticed that our preferred company name indeed had meaning. It fit us perfectly. According to the Urban Dictionary, the word “peecho” is American slang. It is a noun, pronounced pee-cho, referring to a person or thing that is “incredibly cheesy, corny, lame... and trying way to hard to be cool.” Excellent. Just what we needed. We had a mission, a logo and a name. We were ready to conquer the world.
21
Community After the initial set-up of the company, gathering team members was next on our list. If you want to build great stuff, you need great builders. In our case, we needed Java programmers - preferably with experience in the print world, cloud computing and internet application development. On top of that, our programmers should have amazing skills, should be generalists, have loads of passion and be really good looking. That’s a huge amount of different traits for just one human, even after I dropped the aesthetics requirement. Super heroes are hard to find, if they even exist at all. So, why bother? First of all, we didn’t want average programmers. We needed the best. In a small team, you can’t afford to have sub-par members. They drag down the others. Besides that, although I have a tendency to be slightly despotic, I really don’t want to tell people what to do. I prefer them to be an awful lot better than me, so they will tell me how they are going to do their job, instead of the other way around. I am afraid that only the best developers would limit my natural reflex to be annoyingly bossy to a bearable minimum. It took me about four seconds to convince my partner Martijn of that. In fact, our developers should be so good, that they could fix their own problems - in any part of the system.
22
They can’t be specialists. Specialists are lethal in small teams. If you are the only one who knows everything about one thing, the system can’t be managed without you. That’s bad, because then you are the person who will be called in the middle of the night if something goes wrong. You will be a single point of failure. So, we prefer people that know a bit about everything. Every developer should be an analyst, architect, web developer, database engineer and even a graphic designer, too. Uh... let me rephrase that. Forget about “graphic designer” - can I still replace it with “team player”? Thanks. We knew what kind of people we wanted, but we somehow doubted that we could immediately convince them that working at Peecho would be their Most Awesome Life Experience Ever. You see, contrary to the belief, working in a start-up team does not make you rich. There are no leased cars, expensive trips or large bonuses. To compensate for that, we needed people with enough passion to take this abuse. You can’t buy passion. It’s simply not for sale. This is not so bad, since we didn’t have any money anyway. The only way to get passionate people to do stuff is to take the plunge and share your ownership. Apparently, people work harder for their own babies than they
do for money. Or so I have heard from people who know people who have both. Thinking of our giant wish list, we looked at the open source community and decided on a similar approach. The idea of the Peecho developer community emerged: a self-managed group that would operate as an independent, loosely tied group of individuals. This community should own part of the company and be entitled to part of the profits, too. In return for that, each member would be required to invest at least a minimum amount of monthly hours. Most importantly, the community should be able to decide on who takes part in the community. We figured that only the best developers could identify the best of the rest. Normal people don’t even know what these folks look like, you know. It takes one to know one. According to our theory, the only thing we had to do was to find the initial members. Then, quite magically, those members would find us other members. So, it all depended on the brilliance of the first generation. We made a list of potential candidates and started shopping around. A couple of phone calls later, an amazing thing happened: our two top picks agreed to our proposal. Just like that. These guys started working
on what would become our print platform in the cloud. However, the initial team members soon realized that there was more work to do than they could handle on their own. So, they asked two other great developers to join in. The plan worked! Now, the Peecho developer community consisted of four people and really started to kick ass. Within a mere five months, we released the first version of our cloud print platform on a large congress and started winning prizes. But I’ll tell you about that some other time. In the meantime, you can be sure that we have a great team - and they know it, too. After reading a draft of this episode of the Peecho story, one of our community members commented that we should consider ourselves lucky that, even without actually enforcing it, the aesthetics requirement has been met so well. Uhm... I guess that’s a matter of personal opinion. I decided not to comment.
23
The investor game, part I In the beginning of 2010, Peecho development was steadily progressing and all was well in the universe. However, when you are developing a relatively high-tech product, chances are that the time-to-market is larger than your wallet can handle. So, we needed some investment to really cut it. To attract money people, we figured it was time for a glamourous launch and lots of media exposure. Chance helped us out. At an Open Coffee meet-up, we met one of the organizers of TheNextWeb. Old-fashioned Googling will show you that TheNextWeb is an annual internet conference known for its start-up rally - a competition in which start-ups battle to win awards in several categories. We decided to go for it and apply. Peecho was going to compete with the best fresh companies from all over the world! Some glam presentations and interviews later, the grand finale was in sight. The first TechCrunch mention soon followed. There was a slight problem, though. At that point, our software wasn’t anywhere near completion. Luckily, our trusty developers applied some of their last-minute deadline magic. With the conference glooming, our community managed to finish the cloud print platform and a demo application for iPhone to go with it. With just one day to go, we saw the
24
first physical print product come out of a Xerox iGen in the print facility of Greetz - the Dutch greeting card specialist. It was a postcard with a picture of a hideously ugly Peecho easter egg. Until this day, it is handled with reverence. Then D-day came. In the keynote session of the event, Amazon CTO Werner Vogels mentioned our service as one of the “future building blocks of the internet”. On top of that, Martijn’s stage performance in the finale was flawless. To cut a long story short: Peecho beat over 200 other start-ups and won the award for “best business model”
in the start-up rally. Mr. Vogels congratulated us on stage and proved the jury’s point by donating a 10 euro bill - another pièce de résistance in the Peecho treasure vault. It couldn’t have been better. Attracted by all the fuzz, the venture capitalists started moving in. During that summer, we were busy flying over the world doing presentations and assessing term sheets. We spoke with all of them, from local investment companies to international players. It was super useful. VCs are usually very smart people. They ask the right questions and you can learn an awful lot from them. The bad news is that it takes so much time and focus to deal with them, especially if you do not have a lot of experience in the field - like us at that time. While being all excited, we made one capital mistake. The process slowed us down. Instead of running around with investment scenarios on our minds, we should have worried about keeping the hyperproductivity of the days before the conference. Luckily, we snapped out of that temporary paralysis relatively quickly, when our dreamed VC decided to wait a little longer. More traction was needed to prove our business model. We decided to quickly move to plan B. This approach was really
simple. We sat down for about an hour and made a list of friends and acquaintances that we thought would make excellent informal investors - ranging from internet company CEOs to legal experts. We ordered this list by several variables, like “knowledge” and “network” and “do-we-really-want-to-have-thisguy-around-all-the-time”. Then, we started to make phone calls from the top down. We didn’t get very far. About 27 minutes later, we closed our first investment round - possibly breaking the world record for the smallest negotiation-to-paperwork ratio in a seed round ever.
By the end of the year, we held our first shareholders meeting. It was a blast. We love you, guys.
25
Technology Technology is an important part of what makes Peecho stand out. Since you made it all the way to this chapter, let us share some secrets with you. Expensive stuff does not scale We are a start-up, so the most important thing that we considered before we started was simply money - or rather, the lack thereof. Although we required some serious firepower, the fully operational system should cost no more than a few hundred bucks a month. The money issue reached all the way into our technology stack. We could insert something cool here about object-orientation, code compilation or MVC architectures - but in the end, we just needed an extensive development toolkit that wouldn’t cost us. So, the computer science background of our developer community members pointed us to the Java platform. Using cloud computing was a dead give-away, because on demand scaling eliminates expensive overcapacity. Our shortlist contained Google App Engine and Amazon Web Services. These molochs have way more resources and experience with scalability than we do. That’s why we vowed to use as many of their cloud services as possible, rather than building stuff ourselves. However, we needed queueing and relational data storage - exit Google. So, we chose to build on top of Amazon Web Services, using pretty much all of their services. It is cheap and
26
generally reliable. Just to be sure, our entire system runs simultaneously in two AWS availability zones to guarantee maximum uptime. If we go down, half of the world will be there with us. If you are slow, you can’t grow However, technology is not everything. Scalability for small companies is largely obtained by flexibility. Be fast. We try to keep our code simple, so refactoring is easy. As a result, we can afford to only build the bare minimum that is needed without thinking much about the future needs. We can fix those later. If you really want to be fast, short iterations are mportant. On average, we release new software to our production environment about once a week. To keep up with that pace, we organize a monthly dinner. That’s when demos are shown and everybody cheers and applauds for the new stuff. There is a lot of affordable help available to get you up to speed. For example, we develop according to Scrum and Test Driven Development. We use Jira with Greenhopper for requirements management, Bamboo for continuous integration, Sonar for code quality monitoring and Mercurial for distributed versioning. By the way, the large number of Atlassian products is partly explained by their cheap start-up licenses - but it is good stuff anyway.
Sure, but what does the architecture look like? For each customer, the cloud print button is a static file, cached in the Cloudfront CDN. This means that the hits based on views are taken by Amazon and not by us. It is refreshed once a day to reflect price changes and current exchange rates. It gets even better. When a visitor clicks the button, the checkout screens are served as a single static file from the CDN as well. Dynamic data is obtained by a few JSONP server calls only. Being lazy pays off. The server component that is responsible for the dynamic data exchange exists as a separate application on top of our platform. This separation of concerns is meant to minimize the interaction points - to achieve high cohesion and low coupling. The app runs on an EC2 auto-scaling group with a load balancer on top. It stores order data in SimpleDB, the Amazon No-SQL data store that can take a beating. After payment has been confirmed, the print button checkout submits every order to the REST API of our API that our premium users access from their apps, so we only have to maintain a single interface. It is load-balanced to multiple nodes that can scale out automatically. Using the relational database service RDS, order data is kept in two availability zones. Now, this is where the magic happens. The order intake machine writes a ticket to the processing queue. Whenever there are enough tickets available, a new processing machine wakes up, gets a ticket and starts crunching that awful data. We use large
spot instances that we rent by bidding on unused EC2 capacity. When those big guns are done, they store the result in S3 and then kill themselves. So, using queue metrics, we scale the amount of processing power up and down as the number of items in the queue varies. This saves money and prepares us for unexpected outages. Before you do this yourself, keep in mind that each component or module should be responsible for a specific feature or functionality only. A component or object should not know about internal details of other components or objects. Subsequently, the order must be routed to the right production facility by adding a production ticket to the right print facility queue. Using either our Adobe AIR Printclient desktop application or a direct integration, the facility retrieves its jobs from the queue and the product files from S3 file storage. There are no servers involved, so this system is practically bulletproof. In return, the facility can update job statuses by posting messages to the central order status queue. Asynchronously, we read that queue to update our customers about their orders. If all goes well, the product is shipped and delivered pretty soon. After a certain period of time, the product files will be removed from S3 to save storage costs - and everybody lives happily ever after.
27
is a free service that allows anyone to sell professionally printed products from their application or website, and make a profit with every sale. Customers simply sign up, get the print button code and embed it in their website. Visitors can then buy digital documents or images as photo books, reports, magazines, canvas prints, and more. An API is also available. www.peecho.com