Here at Six & Flow, we're big fans of chatbots and conversational marketing. We're convinced it makes a lot of sense to bring the sales and marketing process more in line with the human conversations we have every day on apps like WhatsApp and Facebook, and we have plenty of case studies to prove it helps our clients sell more and sell faster.
But is there a replicable formula for chatbot success, or are conversational marketers making it up as we go along? As conversational marketing matures, it's no longer as simple as installing a code snippet, knocking together some basic scripts and waiting for the sales enquiries to start pinging in.
Nope - in order to be really effective, your chatbot - or rather chatbots - need to be every bit as targeted and singular in purpose as a decent ad, blog, CTA, landing page or other marketing asset. It turns out there is a correct way to build a chatbot - here's how.
First off, chatbots should be contextual. This is a nice and easy principle to grasp, and since day one of conversational marketing we've talked about the importance of matching chatbots to user intent - putting a different welcome message on your pricing page compared to your blog, for example.
However, chatbot targeting can - and should - be a lot more sophisticated than triggering different scripts based on URL or time of day. As a starting point, we'd also recommend thinking about the following:
Obviously, not all of this information is readily available when an anonymous prospect lands on your website. However, that's not to stop you making assumptions about your traffic in the same way you would with other assets (for example, putting a bottom-funnel CTA on a page based on assumptions about visitor intent) - and, as your chatbot starts gathering info, you can confirm those assumptions, segment and iterate.
Next, chatbots should be prescriptive - not passive. Rather than wait for the user to tell you what they need, you should use your chatbot to take them by the hand and direct them towards a solution to their problems.
If it helps, think of how you'd write a blog to appeal to a specific persona and drive them towards a specific action (such as a CTA click). In the same way, you can design a chatbot to target website visitors with specific pains and provide a remedy - whether that's helping them surface an article from your knowledge base, signing them up to your mailing list or booking them a meeting with a sales rep.
Finally, if you want to succeed with chatbots and conversational marketing, you're going to need to be able to measure their effectiveness in driving the visitor towards a single goal.
This is another good reason not to slack off when it comes to making your chatbots targeted, contextual and prescriptive. If you try to accommodate dozens of different needs in a single chatbot, you end up with multiple complex branches and multiple complex outcomes. This isn't easy to build, and it's emphatically not easy to measure success.
On the other hand, if you have (for example) an MQL bot whose job is to book meetings and a customer service bot whose job is to close support tickets, it becomes a lot easier to understand what's working and what isn't.
A chatbot shouldn't be a swiss-army knife - it should do one job and do it well.
Once you start following the above framework, it suddenly becomes a lot easier to build successful chatbots. It also becomes a lot easier to build chatbots in the first place.
Why? Because when you understand that chatbots and conversational marketing should be contextual to specific needs and prescribe specific outcomes, it becomes a lot easier to think about them the same way you'd think about any other asset. It's easier to tie them back to your persona research and easier to scale resource input to results output. And it's easier to build, optimise and measure success.
Want to find out more? Click here to talk to our chatbot and download our guide.