Shopping in Messengers: May 2016
Testing some popular chatbots, what’s working and what could use some work.
Today everybody talks about chatbots — special programs that are integrated into messengers to interact with customers. Why has this become important? Because messengers are where the people are now. Messaging platforms, such as Whatsapp, Facebook Messenger, Telegram, and Wechat, have already outgrown social media sites in terms of audience. People love to chat, to send emojis, and to share photos with friends, so why not create a similar experience for talking with businesses?
The chatbot concept is not new (let’s remember ELIZA from 1966, PARRY from 1972, and SmarterChild from 2000), but it’s only now — given the rapid growth in the use of messengers and the massive development of natural language processing technologies, led by Apple, Google, Microsoft, and Amazon, among others — chatbots could potentially become something big and useful.
As a significant part of the online business world, e-commerce companies can also benefit from this new channel. And since online shopping is so close to my heart, in this post I would like to show you the current state of shopping in messengers.
Overview
It is a very early stage for Conversational Commerce, given that there are several significant limitations in terms of ensuring a good shopping experience. But despite skepticism, thousands of developers are already creating chatbots. Hopefully, quality will rapidly follow quantity. So let’s see how the first e-commerce chatbots are doing.
For this overview, I have chosen four chatbots from different messaging platforms:
- 1–800–Flowers on Facebook Messenger
- H&M on Kik
- Kip on Slack
- JD.com on Wechat
I wanted to add the Spring’s chatbot for Facebook, as well, but at the moment access is closed. If it is actually working, or if you know the reason for the stoppage, please let me know in the responses below.
1–800–Flowers
We’ll start our journey with the well-known bot that was one of the first bots introduced on the Facebook Messenger platform. The 1–800–bot allows the user to search for and order flowers. Simple, as it should be.
Hi, 1–800–Flowers bot!
The first thing I felt while interacting with this 1–800–bot is that it lacks a human touch. As a techie myself, of course I like soulless robots that know the best way to execute a task, but a pinch of “Hi, Ivan! What a beautiful day — you’ve finally decided to join 1–800–Flowers. So what is the occasion today?” would be nice. Instead, the bot immediately asks for my address. This is the same as asking a girl for her number without an icebreaker, yeah? Not a good start, in my opinion.
As you can see in the screenshot above, the 1–800–bot doesn’t give you much space in terms of conversation. I believe the bot’s team will soon make it more interactive.
Featured CBM: Designing a Chatbot’s Personality
Once you’ve entered the delivery address, you can proceed to finding the perfect flowers. The bot will ask for the occasion for the flowers (birthday, thank you, Mother’s Day, etc.), and then you’ll get a 10-item carousel with the product options, followed by a couple further questions about the delivery.
The second thing I found inconvenient was the small size of the product images ; on some pictures, it was quite difficult to identify the product, while on others you could identify the product, but you couldn’t enjoy the view much. When I wanted to tap an image to see a larger version and more details, I was unable to do so.
As you can see in the screenshot above, the 1–800–bot doesn’t give you much space in terms of conversation. I believe the bot’s team will soon make it more interactive. Finally, once you’ve found your perfect bouquet, you will be redirected to an optimized mobile page for entering your bank account details. At the moment, this is not convenient as it should be, so we all are waiting for the payment API to be opened soon.
H&M
The official chatbot of the largest European apparel retailer, H&M, was among the first bots to be presented at the recently launched bot store on the Canadian messaging platform Kik.
Hi, H&M bot!
To be honest, after the robot-like approach of 1–800–Flowers, I wasn’t expecting much from another shopping chatbot, but H&M and Kik have impressed me. Interaction starts with a short and sweet greeting, which is followed by the introductory questions on gender, age range, style preferences, etc. The bot is actually trying to get to know me better, and it does this in a friendly and non-salesy manner, using real-world words and emojis (👍, right?:)
Once the bot gets to know you a bit, it can offer personal style tips with different clothes combinations. I like how the H&M bot displays images — the size is good and there are no cropped elements, so everything is clear and looks great. The only thing that is questionable is whether there should be a solo image or a carousel, as with the 1–800–bot. Hopefully, the A/B tests will answer this question.
When I found the sweatshirt I like, I just tapped the word in the menu that was provided and there it was: a large picture, the price, and the details. Next, the menu provided by the bot asked me to share/save/buy. If you want to buy, you will be redirected to the sweatshirt mobile page on the H&M site. Provide the payment details, push “buy,” and voila!
Kip
Kip is an AI chat assistant for group shopping that works on the Slack and Telegram platforms. Kip aggregates products from a variety of stores in the US, such as Macy’s, Amazon, and Target (80 brands in total, according to the website), and it helps chatters to find whatever they need. It can be used for both group and personal shopping.
Hi, Kip!
The first thing you notice when you interact with Kip is certainly the image. I personally like the cartoon-style character — it looks very friendly, and it invokes an unexpected desire to chat in order to see if it really is so sweet or not. So, I went down this road and started chatting with Kip.
If you want to be on the same page with Kip, just use simple commands with the item titles.
While Kip had a pretty good start, greeting me and responding to my “How are you?” question, Kip then failed to understand my answer and started offering me Swarowski earrings. How about getting to know each other a bit? Or trying to understand who I am and what I am looking for? With the phrase “That is not what I want,” Kip offered me another three products. So, at this stage, Kip doesn’t rely much on the conversational aspect, so there is no point trying to use natural language. If you want to be on the same page with Kip, just use simple commands with the item titles.
With regard to group shopping, which is what Kip was initially designed for, everything works in a straightforward way. You ask Kip for something that you need (a PS4 in our case) and get three options to choose from; you can save one of them to your cart or get more alternatives. The same process is used for your teammates. At the end, you have a cart filled with products for your company, which can then be accessed by an office manager who can order specific items. Good job, Kip!
JD.com
As you all may know, Wechat is a pioneer in terms of mobile commerce in messengers, and at the moment it provides a unique shopping experience to its users. It is not very common for large Western companies to catch up with their Asian rivals, but this is the case here. At the moment, Wechat is the #3 messaging platform globally in terms of MAU (following Whatsapp at #1 and Facebook Messenger at #2) and #1 in terms of revenue. While Western messengers have only recently started to explore monetization opportunities, Wechat already earns money thanks to the “apps within an app” model. In this section, I would like to show you how a lucky Chinese user can order an iPhone from JD on Wechat.
Hi, JD.com!
JD is one of the largest e-commerce conglomerates in China and the world. Users can find JD’s products on Wechat using either of two options: 1) the e-commerce directory of the Wechat Wallet, which is a separate portal where users can pay for services (such as buying a movie ticket, ordering a taxi, or making an appointment with a doctor), or 2) JD’s official account.
So, you won’t find virtual assistants powered by the latest AI; instead, you’ll find light mini-apps with button controls. Wechat, in this case, can be seen as a new browser (or operating system), with the official channels and bots as the new web pages.
But what is the difference between Wechat’s official accounts and mobile sites? On Wechat, you already have your payment and delivery information saved, so if you want to buy an iPhone, you don’t need to provide payment details and worry about data security. This is super convenient.
Although I do not see any radical shifts in terms of shopping experience on Wechat(compared to apps and mobile sites), the key point is that Chinese users are already using it to buy iPhones, clothes, movie tickets, and food on a daily basis. This is something that should be, and soon will be, available to all of us. Cheers!
Conclusion
In this post, I’ve covered my online shopping experience with different chatbots and messaging platforms. My goal was to understand how this experience differs from the app/web platforms and to show you how leading tech companies see conversational commerce as of May 2016.
At the moment, there are two trends in messaging commerce: 1) the Eastern trend (Wechat), where the user has access to products through official accounts, and these products are typically presented through messenger-oriented webpages rather than by chatbots; and 2) the Western trend, (Kik/Facebook/Slack) where the key aspect in shopping experience is the simple (so far) chatbots. I think the optimal solution could be found in the middle, and I very much look forward to see what will happen over the next 6–12 months.
With a strong belief in the future of messenger commerce, my team is also investing energy in bot development. Mr.Chatbot is a one-click solution that allows online shops to promote and sell products with no coding or scripting required. We are actively working with online shops around the world, so please don’t hesitate to sign up for free beta access at http://www.mrchatbot.com/.
p.s. Special thanks to the guys from Bots facebook group who consulted me on Wechat features, especially Clement Tang, KC Alex Liu, and the wise man Vincvinc Vinc.
Featured CBM:
