Unlocking LinkedIn’s Algorithm: What You Need to Know Now
One day in November, a product strategist we’ll call Michelle (not her real name) decided to explore LinkedIn from a different perspective. She logged into her account, switched her gender to male, and changed her name to Michael. This intriguing transformation was part of an experiment dubbed #WearthePants, where women tested whether LinkedIn’s new algorithm showed bias against them.
For months, many LinkedIn users expressed frustration over declining engagement and impressions on their posts. This sentiment followed comments from Tim Jurka, the company’s vice president of engineering, who noted in August that the platform had adopted large language models (LLMs) to enhance user experience by surfacing relevant content.
An Experiment in Perspective
Michelle became skeptical of the performance changes after observing her engagement statistics. Despite having over 10,000 followers, she found that her posts received similar impressions to her husband, who had only about 2,000 followers. “The only significant variable was gender,” she remarked, raising questions about how much impact LinkedIn’s algorithm had based on this factor.
Similarly, Marilynn Joyner, another entrepreneur, reported dramatic changes after switching her profile gender. After altering her profile from female to male, she saw her post impressions soar by 238% in just one day. Other participants like Megan Cornish and Jessica Doyle Mekkes echoed her results, further suggesting a troubling trend.
Despite these unsettling findings, LinkedIn asserted that its algorithms do not utilize demographic information such as age, race, or gender to determine post visibility. They maintained that discrepancies in user engagement could not automatically be classified as unfair treatment or bias.
The Complexity of Social Algorithms
Experts in social algorithms caution that while overt sexism may not be at play, implicit bias could be influencing outcomes. Brandeis Marshall, a data ethics consultant, highlights the intricacies inherent in algorithm design. “The algorithm pulls various levers, which constantly interact to produce results,” she explained.
This means that each change—down to the profile photo and name—can subtly affect how content is prioritized in user feeds. Marshall noted that unexpected spikes in impressions could stem from factors like user engagement patterns or viral trends, complicating the narrative of gender bias.
The Birth of the #WearthePants Initiative
The #WearthePants experiment was spearheaded by entrepreneurs Cindy Gallop and Jane Evans. With a combined following surpassing 150,000, they sought to understand why many women were experiencing decreased engagement on their posts. They instructed two men to replicate their content to see if gender significantly influenced reach. Gallop’s original post reached only 801 individuals, while her male counterpart surpassed 10,000—a stark illustration of the potential disparities in the platform.
Joyner, frustrated by these inequities, expressed a desire for LinkedIn to take accountability for any biases present within its algorithms. However, LinkedIn’s explanations about their models and data usage remain vague, which only fuels further skepticism.
Unraveling Implicit Bias
Despite LinkedIn’s claims, research has demonstrated that underlying biases, attributable to who trained the AI models, could still persist. Marshall pointed out that many platforms may be unintentionally fostering these biases due to the narrow viewpoints embedded in their algorithms.
LinkedIn maintains that the results of the #WearthePants experiment do not demonstrate gender bias. They argue that their systems rely primarily on user engagement and have conducted tests to ensure a fair experience for all users. However, the exact workings of these algorithms remain within the enigmatic “algorithmic black box,” leaving many questions unanswered.
As Marshall indicated, multiple factors could augment visibility on the platform, from the tone and style of writing to the newfound emphasis on engaging with specific posts. Michelle’s experience, for instance, led her to alter her writing style, resulting in a significant boost in impressions.
Addressing Algorithmic Frustrations
Many users across the gender spectrum express dissatisfaction or confusion regarding LinkedIn’s updated algorithms. Data science consultant Shailvi Wakhlu noted that her impressions had dramatically decreased, which she found disheartening after years of active engagement.
While some users experienced luckier outcomes based on topic specificity and audience targeting, a repeated theme emerged: a plea for transparency. As Michelle articulated, people are eager to understand how to navigate the algorithm successfully.
A Final Thought on Transparency
The underlying complexity of social media algorithms makes it difficult to pinpoint the exact triggers influencing engagement. LinkedIn has acknowledged the rise in user activity, suggesting increased competition within the feed could also be affecting visibility metrics.
While insights into what content performs well can guide users, the demand for complete transparency remains a challenging proposition. The nature of competitive algorithms means that companies are often reluctant to disclose too much detail, which could lead to strategic gaming of the system.
In an ever-evolving digital landscape, nurturing authenticity and transparency is crucial. Perhaps the most profound takeaway from this dialogue is that we all deserve a platform that values our contributions equally, regardless of gender.
If you’re passionate about creating engaging content that resonates with audiences, explore how you can leverage your own unique voice in today’s digital sphere. Start the conversation, and let your insights shine!
