You must have heard the hype about chatbots by now. These AI-powered messenger apps are being touted as the next stage of brand-consumer communication. Kiss goodbye to the old days of generic marketing messages because these bots connect retailers with people on a one-to-one basis.
We’ve been talking about the importance of personalised marketing for years already, but this is the first technology to make it possible without building your own data-intensive applications.
So it all sounds very promising. Chatbots are here to save us from marketing ambiguity – great! And eCommerce firms are among the first to bring fully-functioning bots to the masses. So how good are the results so far? Are we talking about a sophisticated platform destined to change the way we sell online or another Siri-like gimmick that doesn’t really do anything useful?
The best of Chatbots (aka the minority)
Okay, so the cat’s out of the bag on this one. The truth is, the majority of chatbots are pretty uninspiring – but that’s fine, because it’s a new technology and we’re going to cut it some slack.
A not-so-riveting conversation with one survey bot in Facebook’s Messenger app, Courtesy of Adespresso.com
There are some examples of chatbots that are doing genuinely useful things, though – even in these early days. Sure, the tech is still a bit crude but there’s promising room for growth in some cases.
True, this doesn’t have anything to do with eCommerce but it proves chatbots can be genuinely useful in the right case. The internet has turned us all into a bunch of whining hypochondriacs but we’ve also got a shortage of GPs on our hands – something needs to give.
The funny thing is, when you see a GP, all they can really do is rule out anything serious and work their way through the list of most likely causes. They’re not specialists; they’re general practitioners – so why not automate as much of this process as possible?
Okay, now for the eCommerce examples.
H&M, via Kik
One of the most exciting names in chatbot tech right now is Kik, which provides a smooth third-party platform for developers to create their own bots. You can see plenty of examples on their website and test the bots out for yourself.
From the examples I’ve looked at, I think H&M offers up one of the best eCommerce experiences on Kik, which looks a little something like this:
The “conversation” is natural enough, without being overly creepy, and it helps you narrow down your options pretty quickly. The platform remembers everything you tell it, learning more about your preferences with every session. So when you come back next time, you don’t have to rn through the same old questions again.
It’s a good enough start – one that says chatbots can provide a real alternative to browsing eCommerce stores. Is it a better alternative? At this stage, I’m not so sure. If they filter results to only show you products that are currently in stock, in your size and matching your other preferences, then we could be in business. I’m yet to see this kind of functionality but I’m sure it’s coming.
Uber, via Facebook Messenger
A much tighter integration than H&M is the Uber bot, powered by Facebook Messenger. Rather than having a full catalog of products to refine for you, Uber has one simple task: get you from A to B.
This simplicity is important because it creates a much smoother interaction between Uber and its users. When your friend sends you the address for their new place in a Messenger chat, you can request a ride in one tap. No opening the Uber app or typing in addresses (unless you specifically need to change it). It’s quick, it’s seamless and – most importantly – it’s actually useful.
The worst of Chatbots (aka the majority)
Rather than name and shame specific chatbots, let’s look at what makes bots disappointing. A recent article published over at Chatbots Magazine collected some of the biggest complaints from millennials about the current crop of chatbots.
Complaint #1: We don’t like talking to robots
There are a number of issues with talking to bots that developers need to overcome. First of all, it’s just feels a bit creepy and there’s no obvious way around this. On the one hand, you can’t forget you’re talking to an algorithm but you sure as hell don’t want them to be so lifelike you can’t tell them apart from humans.
Maybe this will change as we get more used to interacting with chatbots, but there’s a genuine barrier between users and the technology. This is something we’ll probably see with every AI platform (not all bots are actually AI) for some time, until we start to trust the technology.
There’s also the fact typing on smartphones sucks. It’s slow, inaccurate and generally frustrating while voice recognition is nowhere near ready to ease the pain.
Complaint #2: We don’t want to think for you
I’m not sure if this one says more about the state chatbot technology or the human race, but it was raised in the article, so here goes:
Why do we have to think about what to ask your bot? We never had to do that with apps or websites. You should either know what we want by default or make it really easy to get what we want.
This doesn’t mean making a bot that answers any question we might ask. We don’t want to think about what questions to ask anyway! If we absolutely need to text your bot to get something we want, like searching something on Google, okay. But you better not misunderstand us.
I think the point here is that bots need to have a purpose. People don’t want to talk to a computer program for the fun of it. And they certainly don’t want to buy through one when it’s quicker/easier to do it through your website. So tell people what your bot is actually going to do for them and why it’s better than doing it the old way.
Complaint #3: Not respecting user needs
This tends to happen when any major web trend takes off. Developers start building for themselves rather than the users they’re supposed to be focusing on. Once again, I’ll quote from the article so you can take your own interpretation from it:
For those bot developers who make fun of buttons and suggested questions on chatbots: The jokes on you. Because that’s exactly what we prefer.
Your stubbornness to create “AI” in the form of a chatbot is blindsiding you from reality. Oh, so you can create a “real” AI experience?
This is the biggest problem in chatbots as I see them right now. The technology is very limited to what it can actually achieve, but developer ambitions and egos are bigger than it warrants. The fact is users don’t care how intelligent your platform is – they just want it to be quick, easy and effective.
What problem are you actually solving?
If you’re in love with the idea of chatbots then you probably like them for the wrong reasons. Whether you use chatbots or not doesn’t really matter – what’s important is how you solve user problems. If a chatbot is the most efficient solution, then fine. But too many bots serve no real function or actually make the process even longer or more complex.
This is a shame, too, because bots can be great when they’re applied in the right setting. However, chatbots are in danger of becoming the worst marketing trend/gimmick for the next five years because many developers are simply creating them for the sake of it.
Developers think the technology is cool, edgy. While marketers are gushing over the idea of reaching consumers on a one-to-one basis. But what’s the point in reaching out to people individually if you have nothing to offer?
All you’ll end up doing is frustrating people and putting them off your brand and any future chatbot solutions you provide. So always ask yourself: what problem is this chatbot actually going to solve?
Chatbot lessons from the big guns
I don’t want to downplay the importance of reaching consumers on a personal level. This is the future of eCommerce marketing but you need to solve problems first. Just look at the personal assistant apps coming out of Google, Apple, Amazon and WeChat.
Each of these platforms is using bot technology to speed up the search process.
Google’s Allo personal assistant
Why is this important? Because when people hit search engines or personal assistants they have an idea of what they’re looking for. The intent is already there and the platform simply needs to provide the right content in the most user-friendly way possible.
This is where many developers are struggling when it comes to chatbots.
Another example that’s showing promise is Facebook’s integration of Messenger with Shopify. By downloading the free Messenger sales channel, Shopify store owners can update users about their order details and delivery status via Messenger.
Things just got even more interesting too, now that Facebook allows shoppers to browse and buy products from Shopify stores within the Messenger app.
Shopify users can already set up storefronts on Facebook, allowing users to buy from their business page, but now shoppers can reach out to retailers and buy directly from the conversation.
The key thing here is Facebook didn’t go ahead and create one complicated chatbot platform. Instead, it started simple, by solving one problem first (order and delivery updates) and added more features from there.
The future of eCommerce marketing?
So chatbots will definitely play a huge role in the future of eCommerce marketing but I see a rocky road ahead for the smaller brands jumping on board. We’re already seeing too many chatbots built that don’t really achieve anything and I can’t see that stopping anytime soon.
Sadly, this is what happens when technology is opened to the masses like this. Basically anyone can create a chatbot by using third-party platforms and that means the overall standard will be pretty low for now.
This doesn’t mean your chabots need to suck like the majority, though. If you can solve genuine problems with your integrations, this is going to be a great opportunity to offer a new kind of service as the business world goes mad for bots.