How Polywork Could Have Leveraged Data Products to Improve Its Platform

What Is Polywork?
Polywork is a networking platform for professionals who are engaged in different activities. These can be anything from a podcast to a side gig to even a startup venture. Its key differentiator was the opportunity feature, which enabled its users to post potential collaboration scenarios for others to see and engage with alongside the usual social announcement feed. However, lately, the platform has been reduced to a simple resume site, similar to about.me, coupled with a messenger feature.

Despite its questionable move, Polywork may have some potential as it's still a young company. To explore this potential, what better place to look than its data and its business impact?

Data Involved in that platform
The data involved in Polywork is many-fold. For starters, there is the data each user contributes to their profile. Once deployed, this data is fairly static, but a very solid basis for various events. After all, the users of Polywork are there to interact with each other (at least in principle), so there is a strong incentive to consume this data and act on it. This triggers a variety of interactions that generate even more data. From the time stamps to the interest signals (aka "boosts") to all the text in the comments.

Additionally, there are direct messages that can also be a data source, even if they are probably not read by the Polywork team, for privacy reasons. In any case, the platform may still record the who, when, and how, for each one of these messages. On top of that, if a user follows someone else, there is an additional signal that may gradually build a graph depicting the connections and other relationships among the users.

Beyond all this, there is also data related to when a user logs in, what device they use, how long they stay, what external links are clicked, what sites these links point to, etc. In other words, just from the passive activities of a user, there is a large variety of data to be collected, for all users, even the less vocal ones. Naturally, someone inside Polywork is bound to have access to additional data that is collected around the users, so the datasets available are bound to be richer.

But how is this data useful to Polywork? That’s a question every data professional looking at this is bound to ask.

Potential Data Products
Several data products can leverage this sort of data. By the term data products, we mean any kind of application or dashboard that provides useful insights or some kind of functionality that drives engagement and adds value to the end user. For platforms like Polywork that live off engagement, such products are not just nice-to-haves but an essential part of the user experience. Other similar platforms have thrived through the use of such products which help keep these platforms fresh and interesting.

Some examples of such data products are the following:

• Recommender systems (aka, recommendation engine) for users. Although there was already one for various posts, this is a different use case. This product may use user data primarily, though data related to user interactions and relationships may also be leveraged.
• Dashboard of most active users and their most popular posts. Who wouldn’t want to know this sort of information? After all, linking the platform activity data with the users and their posts is as straightforward as it gets, so a product like this would be among the lowest-hanging fruit.
• Potential new groups based on content analysis. Also, recommendations of those groups to the most relevant users. This product would leverage any text data posted by the users, be it articles or comments on those articles.
• Dashboard of most popular opportunities. This would help keep that differentiator feature more relevant. To build such a product one would need platform data related to the various opportunities posted.
• Different tiers of users based on their networks and their activity. Such a product would rely on user data as well as platform data related to the users and their interactions.
• Featured opportunities based on their likelihood of being popular and the number of visits they get. This product would be based on data around the opportunities posted and how the users interact with those opportunities.
• Featured groups based on their growth rate and number of users. Naturally, such a product would leverage user data primarily and how their interactions with the platform involve specific topics.
• Bot-detection feature based on in-depth analysis of regular users. This is a major issue for many platforms, yet one that can be tackled given sufficient data. This data would be related to any telemetry collected around the users as well as the data the various users post.

Naturally, all this is just scraping the surface of what’s possible. However, before a data product is developed it needs to be tied to a particular desired outcome from the business standpoint. Otherwise, it is just an intellectual exercise that may look good but not justify the resources it needs to come about. That’s why the business aspect of data needs to be taken into consideration, both before and after they are developed, as we’ll look at in a moment.

Business Impact of Those Products
The business impact of data products can be noteworthy, enough so to justify building and maintaining them. In this case, there is little doubt that at least some of the aforementioned data products can add business value. How much exactly would depend on how each data product is implemented, how it is used in the platform, and how it ties to the other features of Polywork.

Let’s look at the data products one by one and how they can add value to the Polywork business.

• Recommender systems. These can be very useful for getting more users to interact with each other. If this is a paid feature of the platform, it can drive sales.
• Dashboard of most active users and their most popular posts. This can be an engagement feature that can help users spend more time on the platform and possibly interact with some of these active users more. This can help differentiate the users and create a more interesting pool that can drive engagement.
• Potential new groups based on content analysis. This can aid in content creation since in a group of focused users, there can be found articles and other other content that can help with the site’s SEO. Additionally, this feature can be used as part of the company’s marketing.
• Dashboard of most popular opportunities. This can be a useful feature to drive user engagement.
• Different tiers of users. Depending on how this is utilized, it can be a powerful feature that can help the company engage with different users differently. Some of them may be ready for a pro account, which the company can market to them with reduced costs as they are closer to the target audience.
• Featured opportunities & groups. These are another couple of features that can help drive engagement and make the platform more interactive.
• Bot-detection feature. Last but not least, this feature can help reduce costs by eliminating illegitimate use of the platform by fake users who are there just for the benefits but without any kosher interactivity. By removing bots from the platform, the whole site is bound to be more appealing to new users too while this can also be part of the company's marketing.

Final Words
All in all, there is a lot of potential in Polywork and the data it accumulates through its features (or at least did so until recently). Naturally, this company is just an example of how data can be leveraged to improve the business and make the product more engaging. Materializing this potential may not be a simple task, but it’s something feasible and worth exploring. The only prerequisite in this endeavor is a solid understanding of data work and the ability to reason around it, aka data literacy. Feel free to contact us for more information on this subject.