Marketers in consumer internet space have an ever-increasing expectation to manage organisational goals when it comes to customer acquisition. An effective way to put yourself on a successful trajectory is by spending good time on products you will be winning these customers for, even before you brainstorm platforms, channels, communication, pricing and other marketing mixes.
User mood varies across platforms and they expect different user experience, interfaces (UI/UX) across different products. Tinder and Netflix finished among the most downloaded and used apps in 2018 for the sheer product delights one experiences on these platforms besides their compelling business model.
Left-right swipes, clean and uncluttered interface, easy & seamless browsing across devices, multi-tenancy models are definitely among the WOW experiences combined with the intrinsic human desire to date or watch entertainment.
While we, a bunch of product & marketing enthusiasts, wanted to build a scalable “Party discovery and booking platform“ back in 2016; we also did want to leverage a few new product features and marketing capabilities Facebook messenger bot had to offer.
Why Facebook Messenger Bot?
Facebook launched chatbots back in 2015 and released APIs for developers globally to come and build a customised Facebook messenger bot on its platform. App fatigue was sighted as one of the biggest reasons for engineers around the world to design easier and shorter funnel mobile solutions than conventional applications with compulsive steps of app downloads, sign-ups and on device installation.
Mind you, with current tech ecosystem of 100s of utility, social media and gaming apps, it is becoming quite trickier for end users to manage storage allocation of same on their mobile devices having limited storage capacity.
Contrarily FB chatbots don’t need to be installed locally on your phone, require no additional sign up flow if you are a Facebook messenger user and have almost zero “ready to use” time.
Facebook also decided to make some noise around the launch of chatbots with free and unrestricted use of FB graph APIs, end to end chatbot development learning platform, publishing tried & tested results for growth & marketing community in the form of success stories, free traffic window on some time of the day, and an accelerator program for successfully incubated and shortlisted bots from all over the world.
Our startup Yo Party depended heavily on Facebook and we had built a fully automated, hyperlocal, high inventory party discovery + booking Facebook messenger bot and monetised it from Day 1.
Having zeroed down on the idea of building a high inventory events discovery platform and differentiate ourselves comfortably from other events discovery and ticketing platforms like Nearbuy, Paytm, Insider etc., we struggled for days to find a hack of automating “Events Sourcing” for our application.
Party places and nightclubs don’t have websites or other prominent web presence and crawling was never an easy task to ensure the magnitude of the inventory we were interested in.
It was then we decided to exploit Facebook Graph Event API and the result was a smoothly run, scalably built and fully automated product. This was a key achievement for the product team as it ensured we quickly go to the market and operate with a lean team.
Chatbots as products are naturally a hit with tech-savvy millennials and they being our prime target customer gave us a convincing hypothesis to go this way. Besides, a chatbot for reasons listed above tremendously reduces steps for a potential customer to get started on our product and explore more! This means easier acquisitions and reduced acquisition costs. No?
Articles you may like:
- Everything about Facebook watch party and how to host one
- 10 low-cost content marketing strategy for early-stage startups
- 16 most used free marketing tools for startups and marketers
a) Unstructured chat and NLP implementations
Building unstructured chat soon started looking like a really hard job. People can come and type in anything for your algorithm to return a fitting response. Back in 2016 and even today, developers struggle to totally ace unstructured chat space and artificial intelligence capabilities still would take a good time to pick up for your product even if you are in the right direction.
Building a fully programmed structured chat was however relatively more feasible and we could tot it up with thousands of NLP (natural language processing) phrases. We scripted the entire product flow from onboarding to discovery to capturing potential customer inputs (to show parties that fits the user’s purpose and choice) to purchase to receipt and ticket generation inside messenger itself and yes the customer feedback loop.
While scaling the user base with Facebook ads, a unique user chat (our first key success metric) initially costed us INR 60 for the first 2, 3 weeks and could subsequently be brought down to INR 12-14. This meant that we generating a lead on for less than Rs 15!
Also, we did not have to worry much about humongous app installation costs or traffic churn rate at various steps of the final app install on your smartphone.
a) Deep linking and building search
Users chatting on our platform was one thing and getting them to buy our listed parties was another. When we observed a peculiarly bad conversion rate and a very few buyers, we had to understand why.
A few hours of analysis and customer journey path probes made it clear that people couldn’t land on the same party (after their onboarding) as they had seen in the ads.
They found themselves lost once they landed and could see scores of bookable parties rather than that one party which made them click on the ad. Aha, a problem of deep linking between ads and listed saleable inventory.
It was, in fact, more surprising to find out that Facebook messenger bot didn’t have deep linking capabilities at that time. We had to fix this issue ourselves quickly if we were to expect any encouraging ROI. Thankfully, we could build a “search” feature inside the product that helped visitors to now perform a keyword search on available inventory and pick the party of their choice.
The keywords would mostly represent the artist, venue, theme etc for that party and users well crisply educated on performing this internal search on a relatively newer type of product. The results were encouraging and it’s a product manager’s delight to see users all together perform a new kind of task on their platforms for which there doesn’t exist a similar flow product in the market.
b) Push notifications
It was awesome to find that Facebook messenger bot as product allowed push notifications and this meant anyone having carried out a chat once with Yo Party bot could always be reached back, free of cost.
This also meant you can build a sustainable engagement with your visitors and in turn ensure retention if you use these notifications strategically. Sending time, context, target group relevant messages help in generating better ROI for every push notification you send.
c) Free messenger window
The delight of delights, Facebook will from time to time give your active bot a free slot in its recommended bots strips inside messenger app and voila you will literally see a thousand of folks checking your product for zero cost!
In order to promote bot developers and chatbots as successful new generation product, Facebook exposes new chatbots to millions of FB messenger users and suggests them to try these bots for shopping, gaming, ride-hailing and other digital experiences currently constrained to mobile sites and apps. We remember doing an instant sale of 40k once through that free exposure.
Liked it? Read more:
Looking for an inspiring workspace? Want to connect with like-minded entrepreneurs, marketers and achievers? Download the myHQ app today!
About the author:
This article has been written by Amrit Singh, VP Marketing at Healthians which is India’s largest medical test at home service provider. Prior to his stint at Healthians and Click Labs, Amrit co-founded Yo Party, a party discovery and booking tool.