<?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 Brian Canty on Medium]]></title>
        <description><![CDATA[Stories by Brian Canty on Medium]]></description>
        <link>https://medium.com/@briancanty?source=rss-b11768ff23f5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*EsnHXv2IyxIIQhVu.jpg</url>
            <title>Stories by Brian Canty on Medium</title>
            <link>https://medium.com/@briancanty?source=rss-b11768ff23f5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 14:03:42 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@briancanty/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[How to Win with Conversational Commerce Today]]></title>
            <link>https://medium.com/@briancanty/how-to-win-with-conversational-commerce-today-cf663b98de5f?source=rss-b11768ff23f5------2</link>
            <guid isPermaLink="false">https://medium.com/p/cf663b98de5f</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[sports]]></category>
            <category><![CDATA[concerts]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[bots]]></category>
            <dc:creator><![CDATA[Brian Canty]]></dc:creator>
            <pubDate>Wed, 27 Jun 2018 19:19:20 GMT</pubDate>
            <atom:updated>2018-09-04T23:46:56.702Z</atom:updated>
            <content:encoded><![CDATA[<p>Historians, economists and journalists have repeatedly written about how the advent of the internet had the most significant impact on commerce since the dawn of human interaction, and that impact was further accelerated by the rise of mobile computing. So what is the next step for reducing friction and increasing efficiency in commerce? If we consider the confluence of a few trends — namely the rise of messaging applications, voice controlled devices, and advances in artificial intelligence (specifically machine learning, natural language processing (NLP) and natural language understanding (NLU) — it’s clear that chatbots (like the web and mobile applications) are the next wave of e-commerce. Below are some thoughts on how we got here, and how to leverage these trends to win in e-commerce in the near-term.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wewWm9N1f_vnIltB2OncPg.jpeg" /></figure><p><strong>A Short History of E-commerce Chatbots and Conversational Commerce</strong></p><p>While the genesis of chatbots in the 1950’s has been well-documented, starting with Alan Turing creating the Turing test for intelligence, followed by Joseph Weizenbaum’s Eliza chatbot, with some intermittent ‘hellos’ from projects like <a href="https://en.wikipedia.org/wiki/Artificial_Linguistic_Internet_Computer_Entity">ALICE</a>, <a href="http://www.jabberwacky.com/">Jabberwacky</a>, and SmarterChild (yes, <a href="https://twitter.com/MaddieMidwest/status/481665571776258050">SmarterChild</a>) — the history of e-commerce-focused bots is fairly nascent.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*X0Ma943mn_zvg1WNqJJD_Q.jpeg" /><figcaption>SmarterChild, an early version of a chatbot and one of my best friends growing up ;-)</figcaption></figure><p>The advent of the mobile computing industry spawned messaging applications which continue to skyrocket in popularity and usage. In the hopes of monetizing this growth, the messaging companies (Facebook, Tencent, Slack, etc.) have created platforms for 3rd party developers to more effectively reach users via micro-applications or chatbots. Given the size of the opportunity, developers have explored a plethora of use-cases, like customer support, virtual assistants, social interactivity, marketing/lead generation, and e-commerce. As chatbots quickly oscillated through the phases of the <a href="https://www.gartner.com/newsroom/id/3784363">Gartner hype cycle</a>, many of these use cases were too early (mostly based on shortcomings with natural language processing technology), resulting in naysayers writing off chatbots as a fad. However, in the wake of those conclusions, e-commerce chatbots have continued to make substantial progress.</p><p>The tides of the e-commerce chatbot industry have shifted quickly as developers continue to iterate and learn based on user feedback. Initially, there was a land-grab in an effort to create the ‘one e-commerce chatbot to rule them all’. Companies like Operator, Magic, GoButler, Mezi and Ozlo launched with different takes on a digital assistant to help users make all of their e-commerce purchases — from flights to clothes to transportation and so on — all through one chatbot. While their initial visions were impressive, they have all pivoted away from their original business model for two reasons: predictability of NLP and mismatch in user expectations relative to the technical capabilities of the bot. It’s extremely difficult to train a natural language processing engine if the user requests are disparate. “Book me a car to the airport” followed by “Find me a gray sweater in size medium for under $50” creates confusion for the bot (and sometimes the user!). Even if you’re using an agent/human-in-the-loop, it’s still really hard (given difficulty of training and quality control) AND really expensive.</p><p>We’re now seeing a shift away from horizontal-layer assistants towards vertically specific bots. Different industries, like personalized shopping (ChatShopper, Operator (after their pivot)), travel (SnapTravel, Mezi (after their pivot, which led to an acquisition by American Express), and tickets (<a href="http://www.heyimlea.com/">Lea</a>), have seen a rise of startups filling the gap for users who want a high-touch, low friction way to make purchases.</p><p>If we look to China and the burgeoning WeChat ecosystem, which launched its developer platform in 2013, we get a better look at the potential of chatbots. WeChat has become wildly successful, and is now the main way hundreds of millions of people in China pay for goods and services. Users can buy movie tickets with WePiao, book a ride with Didi Chuxing, and buy dumplings at a restaurant using AliPay, all within the WeChat app. But rather than focusing on using NLP, many of these services built into WeChat are closer to vertical-specific ‘micro-apps’ focused on rapid distribution, as opposed to fully sentient conversationalists. As we look to grow the burgeoning chatbot ecosystem on platforms outside of China, e-commerce bot builders should learn from WeChat’s successes.</p><p><strong>How to Win with Conversational Commerce Today</strong></p><p>Imagine receiving a text from a friend that knows you better than you know yourself, saying “Hey! You would love these Air Jordans! Want me to grab you a pair in size 12?”, and if you texted back “Yes” those sneakers would appear at your front door?</p><p>Curated suggestions from trusted sources are the most accurate, frictionless way to purchase what you really want, with minimal effort. This is the promise of chatbots and applications on messaging platforms — simplicity, personalization &amp; cheap distribution.</p><p>There are three key advantages that e-commerce chatbots currently have over apps or websites in the near-term (hint: one of them is not NLP, at least not at the moment):</p><p><strong>1.) Leverage the platform’s reach for cheaper and faster distribution</strong></p><p>With over 5 million apps in the iOS and Android app stores, the app ecosystem has surpassed its saturation point. To build a standalone app in 2017, you need a lot of money and/or a miracle to reach critical mass, as the effects of the app boom have taken their toll. The average smartphone user now downloads ZERO new apps per month, commonly referred to as app fatigue. Instead, users are spending their time in their top 10 or 15 apps, with most of their time focused on messaging and communication apps. E-commerce companies, even with good retention, are competing for real estate on the user’s home screen with apps like Instagram, Facebook Messenger and WhatsApp. For purchases that the user makes a few times per year, it’s hard to justify needing a separate app, when the aforementioned messaging/social apps are used a few times per day (or in some cases, a few times per hour). Unless you’re Amazon, building an e-commerce app in today’s ecosystem is difficult.</p><p>Given the high costs of acquiring app downloads (in some e-commerce businesses, installs can cost developers between $7–15/user), leveraging messaging apps that are already installed drives acquisition costs to extremely affordable levels (in some cases, $0.05–0.10). Combining the reach of these messaging platforms with certain viral tools that they’ve given to developers, like Chat Extensions (which give users the ability to add a bot into a group chat), allows developers to further drive down the cost and create a competitive e-commerce product with healthy margins.</p><p>Example of a Chat Extension:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FHHETLLZD-tY%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DHHETLLZD-tY&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FHHETLLZD-tY%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/5fc122ffbf8bd22863e3afd07535e162/href">https://medium.com/media/5fc122ffbf8bd22863e3afd07535e162/href</a></iframe><p><strong>2.) Personalize through proprietary machine learning</strong></p><p>Another advantage of the messaging platform is the user’s familiarity with it as a communication channel. E-commerce chatbot developers should recognize this opportunity as way to effectively personalize recommendations to a user, in the same way they’d receive a helpful text from a friend. Instead of spending resources trying to develop complex natural language processing algorithms, investing in machine learning to accurately predict what the user wants to buy and sending those recommendations via text is a much more seamless way to market than email, and it creates a positive feedback loop whereby the user is rewarded for chatting/sharing more information with the bot in the form of more accurate product suggestions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/350/1*gQlC91lRCwtO1VXKsssO6Q.png" /><figcaption>Machine learning recommendations from Lea</figcaption></figure><p><strong>3.) Simplify the user’s life by integrating deeper and deeper into the messaging platform</strong></p><p>Creating unique experiences for the e-commerce customer is a recipe for success, and messaging platforms continue to create impressive tools for developers to build their e-commerce experience into the everyday life of a messaging app user. For example, most applications and websites that sell event tickets are fairly transactional in nature, focusing on conversion at a specific point in the fan’s journey. <a href="http://www.heyimlea.com/">Lea</a>, on the other hand, is designed to accompany the fan on their entire journey — from discovery of the event (through recommending events based on a fan’s Spotify data, social graph, location, etc.), to planning (allowing the fan to select the best seats in a group chat with their friends, using the Lea Chat Extension), to purchasing the tickets (and splitting payment with friends), through attendance of the event (answering questions about the venue, seats, etc.). Doing so makes sense on a messaging platform, but doesn’t necessarily apply to standalone apps or websites that don’t have a messaging component. By allowing users to purchase what they want, while never having to leave the messaging platform that they already spend the majority of their time, chatbots strengthen both the developers’ and the platforms’ value proposition for the user.</p><p><strong>The future of e-commerce chatbots and where NLP plays a role</strong></p><p>At the forefront of the messaging revolution is a battle between each of the major platforms (Alexa, Siri, Cortana, Allo), competing to be the “gateway to the internet 3.0”. In the same way the Google Search is the current gateway to the internet, these platforms are preparing for the next wave of the internet, many believe will be dominated by AI and voice-controlled devices, with messaging bridging the gap between the current state of screen-based graphical user interfaces (GUI) and voice controlled devices. By creating a better device and experience with better general AI/NLP, users will increasingly use these large platforms for a variety of everyday interactions and purchases. These platforms will add value to users looking to complete e-commerce purchases by routing them directly to the best vertically integrated chatbots/micro-apps.</p><p>There are use cases where natural language processing, if used sparingly, can provide a great user experience relative to a graphical user interface. For example, if a user is looking for a certain type of hotel room with certain characteristics, being able to ask in their natural language “Hey SnapTravel, can you find me the best priced room below 14th street for under $200 a night with twin beds and at least a 3 star rating?” is a much better user experience than sifting through a long list of every type of hotel room. The same goes for tickets, for example, asking “Hey Lea, can you find me 4 tickets to the San Francisco Giants game on the home team side in the first 25 rows for under $30 each?”. Surely, there are many instances where natural language processing is more laborious, for example, inputting the fields of a credit card form, but when used properly, can be extremely powerful.</p><p>Building natural language processing tools for your e-commerce chatbot may seem tempting, but for many reasons, it’s not necessary to win in the near term. Furthermore, there are many off-the-shelf NLP solutions that use deep learning across a variety of applications, can handle requests better than a single company ever could by training their bot (Api.ai — owned by Google, wit.ai — owned by Facebook, Watson — owned by IBM, LUIS — owned by Microsoft, etc.). Chat/natural language processing can be used sparingly (or not at all) and the user will still have a better experience than having to download a separate app. In some aspects of the e-commerce experience, using natural language processing (as opposed to a graphical user interface or a pre-loaded response) is actually a worse experience for the user. The best experiences combine natural language processing with pieces of graphical interfaces while leveraging the three advantages of the messaging platforms outlined above.</p><p>Just like we saw in the WeChat example, in the future, a user may say to a friend “Hey, I think we should get a hotel tonight in San Francisco”, and Facebook Messenger suggests the user books with SnapTravel; then the user follows with “it’ll be great to crash after the Jay-Z concert” and Facebook routes the user to <a href="http://www.heyimlea.com/">Lea</a>; followed by “Let’s grab dinner at Outer Lands before the show” and Facebook routes the user to the OpenTable chatbot, for a seamless experience, all done within the Facebook Messenger app.</p><p>With chatbots, while NLP is technically impressive, the way to win in e-commerce in the near-term is through simplicity, personalization and cheap distribution.</p><p><em>In a separate post, I’ll put some thoughts together on how to win with chatbots in the long-term (hint: this does include NLP/NLU). Hit the clap button if you dig :-)</em></p><p><em>If you’re looking for concerts/sports/theater/comedy tickets, give Lea a try — message her on </em><a href="http://www.messenger.com/heyimlea"><em>Messenger</em></a><em>, </em><a href="http://facebook.com/heyimlea"><em>Facebook</em></a><em>, or our </em><a href="http://www.heyimlea.com/"><em>website</em></a><em> and get a discount on your first purchase!</em></p><h4>More where this came from</h4><p>This story is published in <a href="http://blog.usejournal.com/">Noteworthy</a>, where thousands come every day to learn about the people &amp; ideas shaping the products we love.</p><p>Follow our publication to see more product &amp; design stories featured by the <a href="https://usejournal.com/?utm_source=usejournal.com&amp;utm_medium=blog&amp;utm_campaign=guest_post">Journal</a> team.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cf663b98de5f" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>