One concept that I like very much as it is taken directly from product management, is that before building any digital product, we need to consider its desirability, feasibility and viability. With that in mind, I designed two simple but effective social media bots to help increase our marketing efficiency.
The Challenge:
The Startup is new to the market and is trying to leverage social media for growth.
The company has been running its own Twitter and Instagram accounts for some time, but with very small numbers of followers.
Currently, it is posting contents across all of its social media platforms together on a daily basis.
However, since it doesn’t have enough dedicated resources to engage in interactions with its followers,
so this one-way communication on social media is not yielding the expected growth.
The Solution:
After examining the relevant data from all social media platforms, I immediately realized the problem.
While the content posted could be better optimized, but that’s not the main issue since the engagement average falls within expectation.
The challenge that we must solve is how to make our posts reach more people,
and having interactions is one of the proven ways to create engagements and reach for social media.
I tailor made two simple bots for Twitter and Instagram based on the business needs and strategy. For Twitter, I want to connect our account to other accounts who share similar interests with us, and engaged with their posts on a regular basis. Therefore, my bot has two main functions. One, the bot is able to search and follow accounts based on a pre-defined criterion to grow our followings. And two, the bot is able to like and retweet based on a pre-defined criterion as well.
In terms of Instagram, I want to not only like the posts of our followings, but also, I want to interact with people who follow the accounts which is in similar to us. This way, we will be able to grow our account followings organically. Therefore, I built the bot to perform exactly those tasks.
The Execution:
Twitter has a great API for developers to interact with. Using Tweepy, I was able to build this simple bot in a few hours
and perform the tasks as designed. Based on keywords such as “Education”, “Fairness” and etc., it is able to search for accounts to
follow automatically from around the world. It could also be given location limits like “Vancouver”, “United States” and etc.
Also, it is also able to retweet contents based on selected keywords like “teach”, “future” and etc.
The process of building this bot is fairly straightforward, which I won’t expand on it too much here. However, there are two things to consider which have been implemented in my bot but not commonly founded in others.
1. Firstly, I designed my bot to collect all tweets sent by the followed accounts. This is used to analyze what they are promoting and if those accounts fit with our social media strategy. At a much grander scale, hedge funds collect tweets to analyze social sentiments to drive their investment decisions, and our bot does a very basic version of that.
2. The second consideration is that what we should only build in functionalities applicable to the business needs, as this is often not understood very well in the programming community. When it comes to branding, the risk of sending out the wrong message outweights any potential time saved by the bot, so all functionalities must be designed to have the fail safe built in place.
One function that I explicitly built differently than other bots is how retweets are managed. Other bots are just retweeting any tweets based on selected keywords, and that’s not how business accounts are run. The risk of retweeting out some random things from random accounts is way too high, and sending the wrong message would be very detrimental to the brand. Therefore, I specifically built mine based on real-world business applications. The retweet would only be sent out from my bot if it is from a credible/trusted source and it is relevant to our business needs.