Bot Acquisition & Engagement Strategies
A conversation with the bot leaders at Reply.ai, TOPBOTS, Kip and Wade & Wendy
This past Monday, Dashbot hosted a panel of bot leaders at Samsung NEXT in New York City, including:
- Drew Austin, Co-founder and CEO of Wade & Wendy
- Clara de Soto, Co-founder of Reply.ai
- Adelyn Zhou, CMO of TOPBOTS
- Rachel Law, Founder of Kip
They shared their insights and strategies for bot user acquisition and engagement, as well as their hopes and expectations for the future of bots. Here’s a recap of what was discussed. Responses have been edited for brevity and clarity.
1. What acquisition strategies are successful for bots and which have worked for you?
Clara: Leverage existing channels and piggyback on natural behaviors. Large companies have a head start since they have massive user bases to tap into. First, look for user behaviors that are already occurring. Then, figure out how your bot can piggyback off of these behaviors and conversations. This was my approach with KillSwitch, an app I created that removes any trace of your Ex from your profile. We launched on Valentine’s Day and quickly went viral. Traditional press coverage is still the primary form of discovery, so aim to acquire users that way too.
Drew: Clara makes a great point about looking for naturally occurring behaviors and then fitting your bot in accordingly. At Wade & Wendy, we’ve had success by aligning with companies that have a high volume of applicants. They’ve clearly solved the acquisition piece, so we add value by serving as a candidate engagement and enrichment tool. We engage, listen and learn with candidates when these companies don’t have the bandwidth to do so.
Adelyn: Your bot shouldn’t only exist on the messaging platform. Be sure to create a landing page. This sounds so simple, but you’d be surprised at how few botmakers do this. This allows news publications to reference the bot and it also has huge implications for SEO and lead generation. Also, get featured by submitting your bot to Facebook and Slack. Just reach out, you’d be surprised at how impactful this can be!
Drew: To Adelyn’s point, there’s a huge misconception that just leaning on the messaging platform for discoverability is enough. It’s not. More traditional marketing tactics actually work quite well. Focus on building a brand and a community around your product. Focus on creating an experience people love so much that they tell their friends about.
Rachel: Chatbots are much better at acquiring and retaining users than more traditional methods, like newsletters or cold calls. The conversational nature of messaging platforms like Facebook has built brief, consistent and persistent micro-interactions. This is unique to bots and as a result, engaging with a bot is a much lighter burden on the user’s end. For example, it’s easier to click a button within a bot conversation than opening and acting on a company’s email.
2. How do you define success and what KPIs are important to you?
Drew: From a business perspective, we track time-to-fill a role, number of roles a recruiter can actively manage and candidate engagement. From a user experience perspective, we track time spent in chat to gauge candidate engagement, emoji and punctuation use to gauge user enthusiasm and length and depth of responses so we can understand the knowledge extracted from responses.
Time spent in chat has been a really interesting metric for us. We’ve found that for white collar roles candidates are spending 30 min and for blue collar jobs they are spending approximately 11 minutes or less. We don’t know what this exactly means yet, but it’s a new and interesting data point. As we continue to analyze responses and insights gained from chats, we will learn more about the impact of metrics such as time actively engaged.
People generally want to be heard. They want the opportunity to share their story and present themselves in the best and most relevant light for a job they are interested in getting. We look at user experience metrics to help us develop a more empathetic, patient and understanding chatbot.
Clara: What you choose to track is heavily dependent on your industry and your end goal, which should always be viewed through the “Am I providing a great service?” lens. In Reply.ai’s case, it’s actually better to have data that shows users spending less time with the bot. That means we’re solving their problems quickly! Find out what’s most beneficial to your business and then build in these metrics from the very beginning.
Adelyn: I write a ton about bot analytics in TOPBOTS! In fact, I actually wrote the first ever piece on bot analytics (read here). As Clara said, it 100% depends on your type of business. Regardless of industry, you definitely want to be tracking the number of people who unsubscribe. Haha.
Rachel: A good rule of thumb is to use your industry’s metrics. Kip is in the retail space, so we track average transaction value, basket size, the number of participants in a group order, most popular purchases and when people shop most.
3. Platforms are now pushing bots, but are users ready? And if not, how do you encourage them to engage?
Rachel: I don’t believe people are ready for bots. Education is a huge issue. People are still learning how to use bots. Because of this general lack of understanding, you have to figure out where your target users are already communicating and introduce your bot to them there.
Drew: I disagree with Rachel, it’s not that people aren’t ready or comfortable with bots, it’s that people aren’t ready or comfortable engaging with other people these days. With bots, we have the opportunity to rethink what a bot to human interaction looks like. People are ready for automation and decision making support, we just need to educate the public and make sure that expectations and reality are aligned.
Clara: The word “bot” is a dirty word for a lot of people. The biggest challenge is that the consumer doesn’t think about bots in the powerful ways in which they can be leveraged.
4. How do you see the bot landscape evolving and what are you most excited about?
Drew: As bot developers, we should be striving for new levels of personalization. For example, when you go to website or app, the experience exists; it’s laid out in such a way that guides you to behave the exact way the product designers would like you to behave. Chatbots provide a unique opportunity for us to listen and understand the intent of a user, and respond in new and personalized ways, rather than just forcing all users to act exactly as the product requires. This new form of user feedback will introduce a new paradigm of product design insights and opportunities. Future bot focus should be, “How do I tailor the experience to each user?” and “How do I make each user feel understood?”
Adelyn: I’m most excited about bots becoming smarter. The decision trees used by bots today are pretty limited, but in the future, we will have bots that truly understand user intents. Additionally, two big issues are bias and privacy. There are so many biases currently built into the data bots are trained on. We, as a bot community, need to figure out how to address this moving forward. Stringent privacy issues, like those faced by the FinTech space, force companies to take the conversations elsewhere.
Rachel: The biggest issue here is privacy and bias. Like Adelyn mentioned, data collection is inherently biased, which drives bad algorithms. You have to have good data from the very beginning, or you will be training your bot on something biased.
One day, I think we’ll be able to marry bots and have bot children! Seriously, why not?! In the future, I think people will have AI partners, children and friends who are bots. Right now they are tasked with being useful — doing chores, sorting paperwork, etc. As science continues to push further into new frontiers, they can develop empathetic functions and become new and different type of digital toys. For instance in entertainment, bots can take the place of actors, so that people can create their own alternate realities and share these narratives with friends. With small tweaks, you can watch your own living story unfold through the bot actors.
Clara: To piggyback on Rachel’s idea of a bot partner, I’d like to see bots growing into really powerful tools that actually improve lives — not just help solve the small inconveniences that they are capable of today. For example, when a Mom is short a partner, having a bot help her schedule and manage her to do list, would be hugely impactful.
We’ll also finally see the end of hold music! This is a terrible customer experience. Customer service bots, like the one’s we are creating, allow companies to communicate with users in the way they prefer most: instantaneous chat.
Thanks again to Dashbot for hosting! Follow @dashbotio on Twitter and dashbot on Medium.
