What most bots lack

Today’s landscape of bot solution providers and platforms can be overwhelming. If you’re simply trying to understand the nuts and bolts of building a bot, then take your pick.

Once you’ve understood the basics — eg decision trees, webviews, entities, intents and context — then you’ll get to a point where deriving value for customers becomes essential.

The only way to achieve this is to apply a performance driven mindset. After all, the success of the overall market hinges on the quality of the customer experience.

Before delving into the process of unlocking value, its important to set the overall context first. In practice, companies typically take one of the following 2 approaches:

1- Bots are part of an overall marketing strategy to engage (potential) customers on popular messaging apps. Lets call this a channel approach.

2- Bots are part of an overall product strategy to provide virtual assistance inside an existing website or mobile app. Lets call this a product approach.

In this article I’ll focus on the channel approach. We can break this down further into acquisition and CRM.


Say you’re focused on acquisition through Facebook Messenger, here’s 5 things you can do to apply a performance driven mindset:

1Think of bot conversations as landing page experimentation. Similar to landing pages, bot conversations can provide a single, standalone experience. Break this down into 5–10 steps to qualify and capture the lead. Now, each step can be become the basis for A/B or multivariate testing as you try different wording and UI elements to figure out the impact on drop-off.

2 Apply attribution to generated user profiles. When running Ads, its important to understand which Ads lead to the best performance as typically measured by acquisition cost. Each user that engages your bot upon clicking the Ad, will result in an user profile. Regularly, check which Ad IDs and bot conversations were responsible for generating the leads. With this info, calculate your CPL and double down on the best performing Ads.

3 Track sale conversions inside webviews. When optimising for sales, use the bot conversation to end up in a checkout flow (for payment). The checkout flow can be opened through a native webview without leaving Facebook Messenger. When the (potential) customer reaches this step, the Facebook pixel is fired and allows the Ad to optimise further for conversion.

4 Tag important user input. When acquiring user profiles, its important to deepen profiles over time. Basic profile info, preferences, location and order details can then be used to segment users into lists. Subsequently, add Messenger to your CRM strategy and tailor the messaging for each specific user list.

5 Upon drop-off, retarget (potential) customers. One of current approaches to retargeting is to use custom audiences and paid Ads. With Facebook Messenger, you don’t have these costs. Each bot conversation gives you the possibility to retarget (potential) customers within 24 hours after the last contact. Afterwards you have one more shot to trigger engagement for the next 24+1 loop. This opens up tremendous opportunities for further experimentation.

Unlocking such a performance driven mindset is the missing piece to derive value from bots at scale. At Maxwell, we’re excited about building the tools that help marketers do just that. Feel free to reach out to me on asbjorn@maxwell.ai to learn more about how we’ve helped companies like Social Blue, Hellofresh, Catawiki, Zalon, Modomoto/Cloakroom, Adidas, Bol.com and Takeaway.com.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s