Exercise 88 – How would you improve Twitter?

Post and review answers and feedback to answers in the comments section of this post.

See also:

How to answer a product improvement question in a product manager job interview

List of product improvement questions for product manager job interviews

Leave a Reply

avatar
  Subscribe  
newest oldest most voted
Notify of
JC
Guest

I’d start by defining “improve.” Any of the following could be viewed as improvements:
– More profitable
– Higher customer satisfaction
– More active users

Let’s assume that “improve means” higher customer satisfaction. How will we measure customer satisfaction before and after? Customer satisfaction can be evaluated using:
– NPS score
– Stickiness (DAU/MAU)

So we’ve defined what metric we want to improve and how we will measure our success before and after. Now let’s break down the types of users on Twitter by their current engagement levels:
– Power Users – active (meaning they tweet or interact with likes or replies) more than once per day
– Tier 2 Users – active once per week
– Tier 3 Users – active once per month
– Tier 4 Users – active once per year

Let’s focus on Tiers 3-4, as they have the most opportunity to interact more with the platform. Why do they not use Twitter more often? It could be because:
1) They do not find value in the content
2) They do not choose to interact out of privacy concerns
3) They have difficulty using the app

Since we don’t have the benefit of user interviews with this exercise, let’s assume that the biggest problem is #1: They do not find value in the content on Twitter. So how can we improve the content?
A couple of ideas:
1) Make it more personalized. Improve the “Recommended Follow” engine by showing some of their most favorited or retweeted content. Make these recommendations more prevalent than the existing function today.
2) Add “Twitter Recommended” profiles of users that consistently create content that applies to a broad swath of users. Similar to Linkedin Recommends, or “Staff Favorites” at a record or bookstore

Now, which of these ideas do you implement? I’d evaluate based on three criteria: 1) the implementation time, 2) the projected revenue impact, 3) enterprise value. In this case, #1 has an easier implementation since it is enhancing an existing feature, and the revenue and enterprise impacts are the same across both proposals.

So let’s implement idea #1. Test it with a pilot segment of users and engage those users with on-site feedback and interviews to refine and improve. Once these refinements have been made, roll out to a broader audience and measure success using the KPI’s we identified earlier (NPS and stickiness).

Bijan
Guest

Hi JC,
Thank you for the answer. I think the solutions listed here are not novel and this might not work to your advantage. I suggest you list out a few more solutions, including a couple out of the box ones, to help interviewer see your ability to think of fresh ideas. They want to see you come up with big new ideas that are going to help them grow their platform significantly. Check out the answers to some of the other design questions as a guidance. Hope it helps.

Steve
Guest

Great question. Twitter provides a unique platform to engage in real-time with bite-sized information in an open forum; it has become the next-generation wire service. In terms of improving the platform, is there a specific challenge the company needs to address? (Profitability) Interesting, I’ll look at the revenue-side of the profitability equation as I’ve read cost optimizations are mid-cycle at the moment.

User Personas & Pains:
Casual User: Interested in casual consumption of latest news, updates, and ideas. Will participate monthly in a discussion of an activated importance monthly, but generally centralizing on the entities activeon the platform. Twitter may serve as a substitute to phone or email.
Pains -> Finding new content, no influence, confusion over platform value (Platform is rapid distribution, I’m casual).

Power User: Interested in a regular exchange of ideas and discussion. Enjoys recognition and engagement. Enjoys enabling the sharing process to shape a narrative and see what it grows into.
Pains -> Developing a brand without a viral loop, lack of feedback on tweets that are seen as noise, ability to participate in a more meaningful way to effect action.

Key Opinion Leader: Interested in developing brand awareness, influence, and coordinating action to achieve goals, either financial or social. Content creation and editorial skill development are more important to act as curators and acquiring new users. Acquisition is of primary importance.
Pains -> Managing brand risk, improve monetizing/ call to action requests, reaching new verticals of users unfamiliar with content.

Prioritizing Pain Points Revenue Growth:
The success/ failure of Key Opinion Leaders is the most tied into the Twitter platform. Managing success is easier than to create success for someone without success, especially in media (power user); think show ratings and the statistical shotgun approach of pilot episodes. Solving this problem does provide a solutions platform to scale to Power Users nucleate and grow.

There is a measurable value to a product that can target new consumer segments and reducing churn.

Solutions:
1). Peer review platform for content prior to release. The platform can be a focus group of Power Users that provide guidance with a set of metrics to show how well the message aligns with the psychology of the group profile and if it elevates dialog, neutral, or causes conflict. The process can be expedited with ML as we develop more and more user data. In effect this feature becomes an insurance policy to ensure brand is protected. This feature does slow-down the rapidness of distribution, but would be targeted at original content where speed is secondary. Insurance policies are typically valued at 1-3% of value and would be best deployed as a subscription model.

2). Paid access to the feeds of non-subscribers and targeting information for user values to shape content. In the current environment this should be strictly non-political or religious content, but rather offering a platform for Katy Perry or Barack Obama to write an original piece about the state of solar panels and have it seen by the technical community who may engage well with an outside opinion. The product could generate value from a low-cost subscription for the targeting tools and an additional cost for acquired user.

3). Live heat maps for word formation in the tweet to give insight about tone and context. In situations where time is everything, providing a real-time keystroke feedback to make sure tone is properly aligned with intent would add value to ensure messaging is on point. Scalable to include tweet like/ retweet predictions based on past tonal qualities. Most of communication is the author perception and tone so this would lend itself well to predictions. Similar to the insurance model this feature would be in line with an insurance policy at 1-3% of value and perhaps more if it can grow efficacy of message and not just defend.

Evaluation:
I would recommend implementing live tweet tonal heat maps as a first product to test whether there is value in brand protection and possible user growth. The most value to Twitter lives in option 1 about peer reviewed original content, but it is a much larger undertaking before value is understood. The longer feedback loop also hurts viral adoption as the feature is less punchy, despite being more valuable. Option 2 is controversial in the current climate and has been rolled out previously. The unique angle is the develop of customized original content for audiences, but the feedback loop will require a lot of iteration.

Summary:
In summary the best way to improve Twitter’s revenue in a scalable platform is to release tonal heat maps for key opinion leaders to protect their brand and ensure messaging is consistent and possibly increase user acquisition. The model could be rolled out as a monthly subscription with pricing based on AB test results surrouding statistic models of tonal related user churn and user acquisition.

Metrics:
I would measure success of the tonal heat map by measuring percentage user acquisition (new followers, retweets, and positive comments) and churn of users (unfollowing, negative comments) before and after model implementation.

Steve
Guest

Re-reading, I missed:

Metrics would also include new revenue from subscription feature, ARPU, cost to acquire, etc.

Platform is scalable downstream to Power Users who are interested in growing brand. The tonal qualities would help people understand their messaging better and how their user base sees them based on data collected from the KOLs. This could be a much larger market as the number of users is higher, but the pricing would be less to be tailored to the consumer market, but the total revenue per segment I suspect would be higher.

Bijan
Guest

Hi Steve,
Interesting approach. Thanks for submitting your answer. You have a strong argument around why key opinion leaders are important. And I think it’s a good point. I wonder if solution 2 has already been implemented.
One feedback around the evaluation. Perhaps you can list your evaluation criteria to make it clear to the interviewer how you’re evaluating your solutions.

Aj
Guest

Goals of the product: Allow users to post micro-blogs, provide a platform to share & receive thoughts. Social networking.

What’s going well: Very popular, huge adoption, great platform for people to share ideas, thoughts on various topics, great platform for celebrities, public figures to stay in touch with their followers/fan base. Good tool for marketing. High level of engagement.

What are the challenges: Trolling, people claiming their account is hacked so some privacy issues, struggling to monetize, lack of innovation(I could be wrong here) after initial days. Some people find char limit too strict.

Competition: FB, G+, Reddit etc…

Potential Future Goals: Make it a safe environment for people to share their opinion. Increase revenue.

What can be done to address this:
1. To prevent trolling: ML based moderation mechanism. Identify trollers, identify patterns in trolling, identify anomalies and don’t post a tweet until approved by auto moderation system. Another option could be to be let people remove trollers from their followers.
2. Monetization:
– Heavily used by celebrities, celebrities also endorse brands, can we show relevant video ads by celebrities they follow.
– APIs for marketers to min the data for market intelligence. But my hunch is that it’s probably already out there.
– Can offer a Coffee table book which people can get printed or digitally delivered which contains pictures, tweets of their favorite celebrity.
– Can do freemium – pay for extra features such as longer tweets.
– Can provide a way for tweets to appear on TV programs in real time fashion.
– Can provide enterprise product, where people can use to share tweets within the company. Can be used to form communities.

Bijan
Guest

Hi AJ
Thank you for submitting your answer. I think your answer’s structure can be improved by following the format I’ve suggested. Have you taken a look at the “how to answer a product improvement question in a product manager job interview”?
At a high level, I suggest you pick a specific user group, review their user journey and pain points, list out solutions to address their pain points, and finally use some evaluation criteria to prioritize good solutions.
Hope it helps.