Chatbots and Publishers

Do Chatbots create value for Publishers?

Do Chatbots create value for Publishers?

Chatbots is the new hype for agencies and marketers. If you do not speak about or write about it, when meeting the clients to discuss their marketing strategy – you are out. However, do the Chatbot deliver value for the clients and end users? Is it really the new “holy grail”? Will there be an evolution to a new generation of Chatbots, and if so why? 

Let us take a deep dive. 

The first Chatbots and mostly the ones being used right now, requires a dialogue, a back and forth conversation and quite a lot of effort in configuring the rules for the conversation or configuring the “start points” for the AI machine/robot, answering on your behalf. Is this effective, does it give value ? Well, that depends on the use case, if you use the Chatbot to relieve a customer care / customer support situation the solution could help. However, in general, despite what your PR agent has told you, the first generation Chatbots have not been very successful. Even, Facebook recognizes this and David Marcus, Facebook’s Head of Messenger admitted earlier this year “Unripened AI technology and poor understanding of questions led many users to be disappointed with Facebook’s chatbots.” 

What about Publishers, would a Chatbot help them getting engaged readers?

Well, we have CNN and Rappler using the 1st generation Chatbots. Their solutions deliver content. The solutions, however, are hardly user friendly. In order to receive news from CNN, a reader/user needs to make several choices and click 4 times, before he/she finally is ready to receive news. Once this is done, the readers hopefully get the content they want. Furthermore you need to have resources to take care of the questions the Chatbot can not answer on. 

Luckily the 2nd generation Chatbots are now coming, both we at Zummy Direct and Facebook call then Information Bots. These bots will act more like news tickers instead of trying to pretend to be a human discussion partner. They’re not really “chatbots” as much as they are “information bots” as people won’t be having the same back-and-forth conversations as they do with one-on-one chatbots. We have tried the information Bots together with five different publisher clients. 

The mix of Facebook Messenger as the Bot channel and our personalization algorithms has been a success and our Publisher clients experience average OpenRates>80% and average Click Through Rates>60%. A challenge is how to make the readers aware of the Information Bot. We have developed 4 different sign up methods for doing so, since Facebook do not provide any tools for this yet. 

We have started to experiment with what we call the 3rd generation Chatbots. We call them “Actionable bots”. This type of Chatbot is a mix between the Information Bot and the 1st generation Chatbot. In certain uses cases the Information Bot, will transform dynamically to a dialogue based Chatbot “instrumented” with user action buttons. One example of a trigger for the Bot transformation could be geolocation. We mean that this type of Bots could be an app-killer and would thus reduce many of the client specific apps being used today. 

Key Insight

  • The first generation Chatbots had some challenges with unripened AI technology and the users understanding of the Bot questions.

  • There is a 2nd generation of Chatbot, the Information Bot, which requires less dialogue with the user.

  • Using an Information Bot in conjunction with user personalization algorithms for the content be presented, yield quite good KPIs of reader engagement.

  • Which bot type one is to use, should be aligned with the business purpose. Do I want more traffic – Information bot. Do I want to relieve customer support – Chatbot.

  • The number of active users on mobile messaging apps has surpassed social network usage.

  • The latest type of bots are Actionable bots, which transforms to certain purposes dictated by external triggers.