<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Guillaume Boucher on Medium]]></title>
        <description><![CDATA[Stories by Guillaume Boucher on Medium]]></description>
        <link>https://medium.com/@guillaumeboucher?source=rss-2771835d8625------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*SmjKwiX7C0DWFrA0.</url>
            <title>Stories by Guillaume Boucher on Medium</title>
            <link>https://medium.com/@guillaumeboucher?source=rss-2771835d8625------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 21:23:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@guillaumeboucher/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Filtering emails: My very first algorithm]]></title>
            <link>https://medium.com/@guillaumeboucher/filtering-emails-my-very-first-algorithm-ce1cf1220c21?source=rss-2771835d8625------2</link>
            <guid isPermaLink="false">https://medium.com/p/ce1cf1220c21</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[email]]></category>
            <category><![CDATA[algorithms]]></category>
            <dc:creator><![CDATA[Guillaume Boucher]]></dc:creator>
            <pubDate>Sat, 20 Apr 2019 20:30:02 GMT</pubDate>
            <atom:updated>2019-04-20T20:30:02.418Z</atom:updated>
            <content:encoded><![CDATA[<h4>Or how to check if the sender is a personal contact or junk… in an elegant way</h4><p>A few weeks ago, I started a new job at Soul Labs as a backend engineer. It’s a startup who has the goal to create the best AI productivity tools of their class based on mobile, web &amp; desktop apps.</p><p>A few days ago, I was talking to my colleague who created the base of our backend services. He had a challenging problem: we needed to extract contact informations based on the emails our users gave access to and filter which one are personal/relevant contacts in contrast to marketing/bills emails from companies. At first, he create a python script that requested the last 10k emails from a given user and then, parsed them through a filtering system that he created.</p><p>The filtering algorithm he created was pretty simple and intuitive: filter emails based on words that can be found in the email’s subject and body, and filter them again based on the email address it originated from. At first, it seemed to me the best way to go. The only problem is you can’t always add more words and email addresses in a way that’ll filter the emails with a success rate of say… 95% all the time. If a new business comes up and sends everyone 100 emails in the past 30 days, you’ll have to go back in the code and add it as an exception. Not very efficient, if you ask me.</p><p>My dear colleague asked me if I wanted to tackle this complex problem and I accepted. Then I went for lunch. While I was walking outside, I had a <em>Eureka! </em>moment. There might be a better solution so here’s my thought process:</p><ol><li>You receive an email.</li><li>If this email is important to you, say a relative sends you a couple of pictures of their new baby, we assume there’s a very high chance you’ll answer.</li><li>You reply to that email.</li><li>Eventually, there’s an email conversation happening between you and your relative which means you sent AND receive emails.</li></ol><p>Then, if I created a little script that analyzed the emails exchanged, it would look like this:</p><p><strong>Contact</strong>: relative@email.com - <strong>Sent emails</strong>: 10 - <strong>Received emails</strong>: 13</p><p>You might start to see where I’m going with this. If you don’t, here’s an other exampled that I extracted from my own email account:</p><p><strong>Contact</strong>: uber.canada@uber.com - <strong>Sent emails</strong>: 0 - <strong>Received emails</strong>: 76</p><p>Do you start to see the pattern? The most important emails and contacts to me will be the ones I reply to. Since email exchanges is a two-way road, it’s safe to assume the ratio between Sent/Received emails from an important contact will be higher than your basic billing or newsletter email contact.</p><p>But then, how do I really filter the contacts, now that I know the exchange ratio? Do I calculate an average ratio and discard anything that’s lower? Not really. Keep in mind that an average (or mean) doesn’t take into account the extremes and distribution we might have in that kind of dataset.</p><p>Instead, I decided to separate the different contacts in 2 categories: personal, maybe and filtered.</p><p><strong>Personal contacts</strong>: Every contact that is over 0 (ratio)<br><strong>Filtered contacts</strong>: Every contact that is under 0 (ratio)</p><p>With those two categories, I’m able to get a better understanding of how a user interacts with someone else. If you exchange emails with someone, the exchange ratio will be over 0 whereas if you receive a billing email from a service company, you’ll likely never gonna respond to them.</p><p>That’s my solution. But it’s not perfect. Here’s one the drawbacks:</p><p>Say you contact a company once for X reason. You send an email, you receive one and you reply again. The ratio will be 2:1. Then, you’ll have a new contact that’s not really personal but registered as one with that algorithm. Unfortunately, I don’t see any other way to mitigate than adding more filtering layers to have a better general context of your communication process.</p><p>Since I haven’t tested this methodology on a big scale dataset, I’m still trying to perfect it but isn’t that the fun in this? 😃</p><p>— — — —</p><p>If you want to see a follow up of this article where I show you results from my experiment, give me a few claps, share and follow! This is the first article of many where I’ll cover the different tools I’m creating!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ce1cf1220c21" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quick thinking about fragmentation]]></title>
            <link>https://medium.com/@guillaumeboucher/quick-thinking-about-fragmentation-2d4be3675371?source=rss-2771835d8625------2</link>
            <guid isPermaLink="false">https://medium.com/p/2d4be3675371</guid>
            <category><![CDATA[services]]></category>
            <category><![CDATA[ecosystem]]></category>
            <category><![CDATA[google]]></category>
            <category><![CDATA[apple]]></category>
            <dc:creator><![CDATA[Guillaume Boucher]]></dc:creator>
            <pubDate>Sun, 18 Nov 2018 11:15:29 GMT</pubDate>
            <atom:updated>2018-11-19T04:56:59.624Z</atom:updated>
            <content:encoded><![CDATA[<p>Someone asked me recently if YouTube Music would take off. My answer was no. The reason? Pretty simple.</p><blockquote>Google doesn’t know how to create ecosystems.</blockquote><p>You might ask why am I saying that since they have the Google Play Store, YouTube, Android… from a customer perspective, yes they have an ecosystem. I can’t agree more on that view. The problem lies within. For the sake of it, let’s go back a couple of years ago.</p><p>Back in 2000-ish years, we had MP3 players. I digress but I had one and thought it was pretty fabulous. Couple of songs in your pocket and it was a lot of fun. The iPod and iTunes came around with a simple logic. What if you could have a boatload of songs and buy it legally? That was quite a stretch. A lot of people bought into it (myself included) and it worked. In the end, it was a successful business and still is since Apple Music surpassed Spotify in paying customers. Oh yeah, they managed to change their own system. But the real question is why is there so much paying customers for Apple Music?</p><p>It’s quite simple actually. Apple has been able to create a whole hardware ecosystem and understood how much software would be a major player in business for the years to come. They knew devices would eventually look alike very much and the reason customers would pick a device over one is the services. Integration. Seamlessness. Their customers bought into it and data proves it. So how does this translates to the way Google does it’s business?</p><p>We have to think about Google in a very simple way. They are a giant lab. That’s it. They try a lot of things, they fail a lot but in the end, they learn so much and that’s the way Google works! They aren’t in the entertainment business! They are in the upper business! Search, interest and much more. And I think this is the reason they have failed immensely in creating a cohesive ecosystem for their users. They are deeply rooted in user prediction. How we interact with stuff. Not how we interact with it. So for a minute, let’s compare it to Apple.</p><p>Apple decided to create hardware around software so it could best use their users. iPhone with apps and web services, Macs for productivity and professional uses, iPads for content consumption. The reason they create certain services around these are based on data that proves there is a market they can improve.</p><p>On the other hand, Google is just trying stuff. And that’s okay! But the reason they haven’t being able to compete with iMessage, let’s say is because they don’t have the ability to control everything that’s happening with their OS.</p><p>The fact they don’t have a clear strategy on how to merge all these services into one cohesive subscription or whatever, is a proof of how Google works. And that’s fine with me.</p><p>Google is a knowledge company. They know a lot. That’s how they work and that’s why I think their future should be only based in smart home devices (we can see it with their Pixel Stand). But they will never be able to match the way other companies (Microsoft included) are trying the merge the way we interact with services and devices. So in the end, the best will come out of the bunch, even though we don’t necessarily like it that way.</p><p>So now, let’s come back to YouTube Music. As I’ve stated earlier, the biggest problem is cohesion. They have so much services, you can’t really understand how to use it. As a consumer, it’s a headache and this where the problem resides. Some people may say it’s about a user friendliness problem. No. It’s about user comprehensiveness. Make someone understand how something works is not that hard. You just need the approach that’s necessary for someone to really agree with how you do things. But since Google is so focused on how things work, they will never be able to create value for their customers. And for me, that’s okay.</p><p>They are doing a great job at other things.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2d4be3675371" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Shipping Challenge]]></title>
            <link>https://medium.com/@guillaumeboucher/the-shipping-challenge-32cd80a516e8?source=rss-2771835d8625------2</link>
            <guid isPermaLink="false">https://medium.com/p/32cd80a516e8</guid>
            <category><![CDATA[shipping]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[easypost]]></category>
            <category><![CDATA[ecommerce]]></category>
            <dc:creator><![CDATA[Guillaume Boucher]]></dc:creator>
            <pubDate>Sun, 30 Sep 2018 20:01:01 GMT</pubDate>
            <atom:updated>2018-09-30T20:01:01.400Z</atom:updated>
            <content:encoded><![CDATA[<h4>The last mile isn’t the worst. It’s communication.</h4><p>When you talk to customers about shipping, the first word they might say about it is stress. Why is it the case?</p><p>When I was working on Gears (the e-commerce platform I decided to create a couple of years ago), I was thinking about the shipping challenge. From my own experience and reading about it, shipping is the worst thing in online commerce. You never know when your order will come, the state it will arrive in and more. So I decided to think about how I could improve the way it would be handled with my platform. At the time, I decided to do a little experience of my own: order something online (it could be anything) and write down everything about my experience. So I did.</p><p>I came across a product online that was intriguing to me: you could make pastas with a zucchini. I ordered it and was attentive to all of the process: add to your basket, finalizing the order, finding my receipt, etc. If you’re big on e-shopping yourself, you know the process is pretty much the same everywhere. You can’t reinvent the wheel and it’s fine like that. But I started to notice a few differences: some websites give an approximative ETA of your package before you conclude your purchase and some don’t. I wrote that down. Then the long and difficult process of waiting for my order started. To check where my package was, I needed to download an app that manages parcel tracking, put in the reference number and buy a membership of one year to get notifications when the state of my parcel changed during shipping. I can tell you it was tedious. I finally received my purchase one week later with a new app on my phone, a membership I never used again and bit of frustration.</p><p>Then I realized something.</p><p>The problem with shipping is not logistical. It’s more how can we communicate to a customer (who paid hard-earned money for something they want) where their order is and when they will receive it. We’re not always at home, looking at the front door to wait for the delivery guy. So if we have a little bit more information, we can decide which shipping option we buy and where we will receive it.</p><blockquote>The problem with shipping is not logistical, it’s information to the end customer.</blockquote><p>I searched for tools online to get detailed shipping information on the customer’s orders. I found Easypost. They’re a new company focused on an API giving you every bit of information about tracking a package (and much more!). Here’s a list of capabilities that Easypost provided to my website:</p><ul><li>Real-time parcel price based on weight and package measurements</li><li>Estimated delivery days</li><li>Real-time updates on package tracking</li></ul><p>The last point is very important. It’s what helped me put in place a way to help my customers get the information they need as soon as it’s available. Now, I needed to think about the communication channels I would use to serve that information in the best way possible.</p><p>Normally, for that kind of feature, you would use email but since people don’t really read their emails because they receive too much of them, I decided to take a different path. Customers could choose between SMS (provided by Twilio) and connect with their Facebook account to receive the notifications through Messenger.</p><p>I can tell you it worked flawlessly. You would get notifications all along the package route and the day before the delivery, you would get one telling you the approximate hour of arrival (thanks to Easypost’s API).</p><p>In the parcel shipping world, we always talk about the last mile problem being a burden for everyone using shipping for their business. But really, customers don’t really care about it. They simply want enough information so they’re up-to-date with whatever transaction they’ve done. This lesson is good for every kind of business. Before thinking about bigger problems, break them down and work on the smaller parts that will have immediate and positive impact on your customers.</p><p>My problem was not shipping itself but more how I could handle it so I was less dependant on someone else to partly solve a problem that I can’t handle all by myself.</p><p>When you’re dealing with third parties in your business and there’s not much you can do about their own way of dealing with a problem, find the simplest solution that’ll be good for your customers… and don’t forget to voice your complaints to said third parties! Feedback is really important.</p><h4>Thanks for reading me!</h4><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=32cd80a516e8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The best marketing is the cheapest one]]></title>
            <link>https://medium.com/@guillaumeboucher/the-best-marketing-is-the-cheapest-one-b2089482f31c?source=rss-2771835d8625------2</link>
            <guid isPermaLink="false">https://medium.com/p/b2089482f31c</guid>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[lean-startup]]></category>
            <category><![CDATA[communauto]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Guillaume Boucher]]></dc:creator>
            <pubDate>Fri, 21 Sep 2018 00:34:59 GMT</pubDate>
            <atom:updated>2018-09-21T00:39:39.665Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_-hGJQ5t8d4FNzzihP4HQA.jpeg" /></figure><h4>If you’ve got something people want, you won’t need to pay for ads</h4><p>Through my reading and analysis of different companies, I’ve found a interesting trend regarding marketing and growth. I like to watch some episodes of Shark Tank from time to time (you can judge me) and I’m always amazed at how much startups can spend on advertising. I think this is so wrong. Let’s dive right into it!</p><p>A couple of years ago, I started to work as a call center agent for a Quebec company called Communauto. The idea of the service was quite simple. You get a membership with a plan, you rent cars and you get charged for your usage based on your plan. The goal is to replace a personal car with a shared one in order to help your finances and help our environment. Cool, right? Absolutely is. When I was in training, a day was taken to take a look at the history of this thriving business. One thing struck me during that day. Communauto wasn’t actively spending on marketing and yet, their growth rate was quite impressive for a 20-some year old company. In 2013, they had around 30 thousand active users and growing at a rate of 10% each year (it was previously 30%). So how they achieved this? Pretty much all with word of mouth. Not bad at all!</p><p>For the past two years, they started to throw a little money on ads to help get the word out a bit more.</p><p>People are attracted to products or services they find useful in their lives. This is why laptops are became more popular than desktops or Facebook Messenger is used by more people than Skype. They will organically move towards something they find to be best suited to their needs. In a business context, it’s really important to understand that. And Benoit Robert (the CEO of Communauto) understood that pretty well.</p><p>He knew car ownership will eventually come to an end and people needed an alternative. The environment trend would be a a good factor of people switching to his service as well. Basically, he had a good recipe for success and to this day, his company is growing and there’s a clear indication nothing is slowing. For that, I have to give him a round of applause.</p><p>So why do I say the cheapest marketing is the best one?</p><p>Many companies have failed in the past even though they thrown huge amounts of money on advertising and getting a lot of attention. So why do they keep failing? They simply don’t have something people need and they never realized that. They never <strong>pivoted</strong> in a new direction.</p><p>This is exactly why I say the best marketing is the cheapest one. If you bring value from your service or product, people will be attracted to it naturally and you won’t need to spend a lot nor have to make a lot of efforts in order to have recurring and paying customers.</p><p>But now, how do you make sure you can bring that kind of value for your targeted audience? Demonstrate your idea in it’s simplest form possible. With the tools available today, you can cheaply make a video demonstrating some features of your product or publish a website with a couple of screenshots of your app and get people to enter their emails to see if they’re interested.</p><p>You will gain a lot of insights about how your customers feels of your product and then be able to take a decision based on that data. Is it a good idea or should I change it? In the end, your product is not the most important thing. It’s your vision. Your product is simply a way to convey your vision and it can be changed to really appeal.</p><h3>Thanks for reading me!</h3><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b2089482f31c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Eric Ries Taught me]]></title>
            <link>https://medium.com/@guillaumeboucher/what-eric-ries-taught-me-bb9672a12af?source=rss-2771835d8625------2</link>
            <guid isPermaLink="false">https://medium.com/p/bb9672a12af</guid>
            <category><![CDATA[mvp]]></category>
            <category><![CDATA[eric-ries]]></category>
            <category><![CDATA[perfectionism]]></category>
            <category><![CDATA[lean-startup]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Guillaume Boucher]]></dc:creator>
            <pubDate>Thu, 13 Sep 2018 22:09:03 GMT</pubDate>
            <atom:updated>2018-09-13T22:09:03.300Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DOwtmMDN85aUPitEQZ4FqQ.png" /></figure><h4>Or why perfection isn’t always the best way to create</h4><p>Some people know, some people don’t but for the past few months, I started to build a startup. Yes, another of my crazy ideas. But hear me out.</p><p>Two years ago, I decided I wanted to create my own e-commerce platform based around the idea that brick-and-mortar shops weren’t on their way to the graveyard but simply lacked a way to integrate their current business with the correct technology to empower their customers and create value for everyone. I was taking on a project that was gigantic (and probably crazy… again). So I decided to put myself through days long of coding and creating something that would normally require a little team of experienced developers.</p><p>After 18 months, I gave up because it was simply too big of a project for a single guy in his sweatpants and his laptop. I lacked a real structure (and strategy) to help me get my vision to something tangible. I then took a break from coding having moved to a new town and I started to reflect on why I really gave up. What was really the problem with the execution?</p><p>I came to the realization that the project nor my vision was the problem. It was something way more profound.</p><p>I consider myself a perfectionist. When I want to create something, I want it to be the most beautiful, practical and useful thing packed with features. I really do aspire to perfection whenever I’m doing a little personal project, creating a tool for myself or simply having a startup idea. <strong>But perfection is a double-edged sword and I came to realize that the hard way.</strong></p><p>Too many features, too much tweaking. I swear I have changed the position of my main page logo probably a thousands time to get the perfect alignment. I have redesigned the main page entirely 10 times and to this day, I still take a look at it from time to time and would change it again (if it wasn’t from the fact I just don’t want to work on this anymore). I was just straining myself when I shouldn’t have.</p><p>This always plagued my projects but since a couple of weeks ago, it’s not anymore.</p><p>I went to my parent’s lake house not long ago and stumbled across a book called <strong><em>The Lean Startup</em></strong><em> </em>by <strong><em>Eric Ries</em></strong>. Curiously I asked my father what was that book about. Simply put, Ries provides a few methods to create a sustainable business using techniques such as Persist or Pivot, the Minimal Viable Product (or MVP) and much more. Basically, it’s applying the scientific method to create a startup that won’t fail if you’re looking at the right data, being able to realize what’s working and what’s not as well as coming with the simplest product to iterate faster. It’s greatly inspired from the Toyota Method, the Kaizen philosophy and such.</p><p>I don’t think I need to tell you how fast I read that book. It was a real eye-opener. Of course, the book is primarily centered around software businesses but it’s certainly possible to apply the Lean method to almost any types.</p><p>Let’s come back to the fact that I have (another) crazy idea. As you can probably imagine, it’s centered around a mobile app. As I said earlier, I started this project a couple of months ago. After I read that book, I came back home with a totally different way of viewing my app and how I should get it out in the world.</p><blockquote>Something happened the night I came back I would never thought I could do: I started to delete features and parts of my app. And I thought I was crazy.</blockquote><p>One thing the book taught me that really had a big impact on my own reflections was the <strong>Minimal Viable Product</strong>. It’s a way to create your product (in this instance my app) in a way to only have the most basic features and the least perfect product you can provide. We don’t care if it’s full of bugs. It just needs to work. That way you can rapidly iterate on it or just take a new path after analyzing how well you’re doing (backed up with good data).</p><p>I started this new project being tangled in the same perfectionist way I would always do and coming back from this trip, I knew it wasn’t right. And now, I’m on my way to have something that will be way faster and easier to launch.</p><p>The message I’m trying to convey with this first and rough article is somehow a lesson about perfection. Through that experience, I realized in a startup context, perfection is less necessary than your ability to get it out as fast as possible to start analyzing and collecting data to <strong>Persist</strong> on what you have or <strong>Pivot</strong> in a new direction to make sure you get your vision out the right way.</p><blockquote>Perfection is less a way to create or think than simply the byproduct of iteration and evolution.</blockquote><p>If you’re interested in creating your own startup, I can’t push you enough to read that book. It gives a whole new perspective on how you can make your vision and your ideas possible in a way that’s sustainable for anyone involved in your new venture… as crazy as it might be!</p><p><strong>Bonus: </strong>Another great book that I can recommend is<strong> <em>The Innovator’s Dilemma</em> </strong>by<strong> <em>Clayton M. Christensen</em>.</strong></p><h3>Thanks a lot for reading me!</h3><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bb9672a12af" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>