Author’s note: I wrote this before Localmind was acquired by AirBnB and it was originally supposed to be a part of a book that I contributed to called The Everything Guide to Mobile Apps. Because of that acquisition, this case study was not included in the book so I get to share the lessons with you here.
Technology supporter and vocal pundit Robert Scoble called Localmind the “best SXSW app” for 2011. High praise for a company launching their very first app at that festival in the spring of 2011. Coming from Robert Scoble also meant massive exposure in a community of early adopters.
The app, built for Apple’s iOS and Google’s Android operating systems, allows users to ask questions of people at specific locations in real time. If you are thinking of heading to a popular restaurant and would like to know how long the wait is for a table, find a Localmind expert at the restaurant and ask them directly. If you are new in a city and are looking for things to do – ask a Localmind expert for their advice. This service has been described as a cross between the popular Foursquare check-in service and Quora, the question and answer website.
Since the praise from Robert Scoble, they have refined their offerings based on tremendous user feedback and, in so doing, have made the product more valuable. But it wasn’t easy.
They faced some serious challenges after launching the first version of their app – challenges that required them to reengineer certain aspects of their app in order to increase usage and value for their users. Doing a deep dive into the usage data they were collecting, the most obvious area they needed to improve upon was the frequency of use. Asking venue-specific questions was not as normal of often a practice as they had anticipated so they needed to expand their offerings to include larger surface areas – neighbourhoods, cities and countries. They always knew it was a move that would eventually need to be made and their data showed that users were already looking to connect with broader domain experts so they updated this feature in order to increase the engagement rate for their users.
Next, they needed to tackle a smarter algorithm for identifying their “experts” in the field – those people that will be answering the questions posed. What made a person an expert and what was the backup plan should they not be able to answer a question in a timely matter. The founders of Localmind understood that in order to get deeper loyalty and engagement (thus app use) they needed to provide a flawless experience – meaning 100% of the questions needed to be answered within a specific timeframe and in a helpful way. Improving the answer rate and time, while ensuring their experts received enough questions to keep them active within the system, would create continuos and consistent use. This is the panacea for mobile: Being the defacto app that always delivers.
In building Localmind, a few key “Do’s” and “Don’ts” emerged to help ensure deeper user engagement:
– Earn trust. Trust to Localmind meant successful answers to questions about locations from experts that know their stuff. It is a simple equation that sees an expert answer a question about a location in a useful way and in a timeframe that allows the answer to be useful to the asker.
– Make it human. Any hint of machine to human, automated bot-like behaviour and the trust is jeopardized. The founders noted that they are seeing a decrease in machine “search” (think Google) in favour of niche find like using Yelp! or services like Apple’s SIRI. The idea that humans like to ask of and receive answers from other humans is critical to the experience they provide.
– Hide from personalization. People like to see pictures of other people – especially if it is supposed to be a social network connecting humans to other humans. A simple tactic Localmind used was to add actual photos of the experts and those asking the questions. Don’t hide the human aspect of the service or product – if there are humans involved, showcase them!
– Bribe users – because you can never pay them enough! Offering payment for something like answering a question is a double-edge sword. On one hand, the offering seems like a great incentive for people to answer questions – pay to answer model. On the other hand, what is the appropriate value of for the answer and will that be insulting to the expert? Probably. Engagement happens out of a desire to help and contribute. Putting a value on that sets a level of expectation and, often, disappointment. Offer rewards but first, find the key to engagement without.
Data plays a considerable part in ensuring the app remains relevant so collecting the right data – focused on the right metric – is the most important factor for success. Determining and focusing on that metric – for Localmind it was time to answer – helps to decide the features to be added (or removed), the algorithms that need to be tweaked, and the key(s) to customer acquisition and retention. Only collect the data that is needed and nothing else – data spoils quickly so if it won’t be used or is in support of the main metric, don’t collect it.
Refining is a considerable part of the app economy – this is not a “set and jet” business. To do this right, to build a business from an app, there needs to be some care and tending to ensure it is still relevant, still engaging and still providing value. If tended to, the outcome of this will be a strong and loyal user base. If ignored, the impact could be severe brand erosion and irreversible harm.