Blogs/Innovation/Discovering Career Paths: Unveiling the Dynamics within Organizations
Cover Image
Author Picture
By Martin GonellaVice President, Product Dev at Taller

Discovering Career Paths: Unveiling the Dynamics within Organizations

10 MIN read

Discovering Career Paths: Unveiling the Dynamics within Organizations

Have you ever been curious about the different career paths available within a company? How do individuals advance from one role to another? Is transitioning to a different career path within the same organization possible?

These are questions that have long intrigued professionals and companies. In this article, we introduce an innovative process that automatically deduces career paths within organizations, revealing the dynamics and potential for growth within a company.

It is crucial to understand the different career paths available within an organization. Doing so can offer valuable insights into the workforce, and help pinpoint areas for improvement. By analyzing the movement between positions, we can identify roles that may be difficult to fill, and assess the feasibility of career changes within the company. For example, we can determine the likelihood of a designer moving into a software developer role, or whether a software developer can aspire to a management position.

Introduction to Career Paths

Planning a career involves creating a long-term strategy for moving into specific jobs that lead to a desired occupation. This approach enables individuals to take control of their career development and align their goals with opportunities presented by different roles within a company.

For an organization, career paths are powerful tools to boost employee retention and engagement. Well-defined career paths motivate employees to strive for growth and progression, contributing to the company's short-term and long-term goals and reducing turnover rates.

However, many companies have limited knowledge about the diverse career paths available to their employees, often focusing on just a few well-defined routes while disregarding the bigger picture.

Lack of career growth opportunities can lead to dissatisfaction and eventual attrition. Employees who do not experience timely salary increases or title changes are more likely to seek opportunities elsewhere. Employee turnover can be costly for organizations, necessitating the replacement of departing individuals.

In the current era of remote work, companies that offer clear career growth paths are desirable to employees, and implementing robust professional development programs can save companies significant time and resources.

Our tool helps companies and employees solve these frustrations by automatically uncovering potential career paths within an organization. Utilizing publicly available information from employees currently working at the company, it constructs a directed graph that models their behavior and transitions.

To comprehend the possible trajectories that employees can embark upon within a company, it is necessary to explore not only vertical growth within the same domain but also potential career changes. For this, we use our xx.

The insights derived from this analysis are invaluable, enabling companies to identify positions that are difficult to fill and require external recruitment. Conversely, it also reveals positions that can be filled through internal promotions, thereby nurturing talent from within.

A visual depiction of what is being written about

Fig. 1 displays the key components of our career path estimator. When estimating career paths, we focus on specific details from input profiles, like job titles, raw descriptions (if available), and role duration.

When dealing with raw data, people describe their positions and roles in many different ways, resulting in various possibilities for a single role. Sometimes, multiple roles are combined into one job title. To address this, our title normalizer and seniority normalizer are crucial during the pre-processing stage, as shown in Figure 1.

Lastly, the career path engine analyzes a selection of normalized profiles, identifying typical career paths between two positions and considering any anomalies in the flow to spot talent management concerns.

Title Normalizer

To normalize position titles, we first remove stop words and special characters. Then, we use a dictionary of common phrases or n-grams found in lists of normalized roles. These roles include Backend, Frontend, Mobile, DevOps/Cloud, QA/Testing, Data Engineer/ETL, Data Science, Data Analyst/BI, Graphic Designer, UX/UI, Non-IT, Executive/Directors, and Management.

After cleaning, we may encounter two problems: A) a title that matches multiple roles and B) a title that doesn't match any phrases in the dictionary.

For A, we only analyze the subset of already matched roles. For B, we compare against all possible roles. If the similarity between the title and a normalized role isn’t high enough, we discard the profile. This is because misclassifying a title has a more negative impact than missing a sample.

Sentence Similarity

Our proposal to measure the similarity of job titles against normalized roles involves creating a dictionary of sentences that describe each role. We obtain typical sentences for each role by analyzing frequently found sentences in job descriptions. We then use custom semi-supervised multi-class sentence classifiers to classify the remaining sentences.

Seniority Normalizer

The Seniority Normalizer is similar to the Title Normalizer as it involves pre-cleaning job titles. Our next step is to search for keywords in a reference keyword dictionary to normalize seniority levels. The seniority levels we consider include Intern, Junior, Semi-senior, Senior, and Lead.

If we cannot find a match or multiple matches, we use a disambiguation process that estimates seniority based on the time spent in the role.

Career Path Engine

To determine suitable career paths, we use a directional graph with normalized positions as vertices and edges with weight strength. To maintain consistency across analyzed companies, we use a fixed set of normalized positions for the graph.

We then use Yen's and Dijkstra's algorithms to find the K-shortest paths between each pair of normalized positions while disregarding paths with downgrades.

Experimental Results
A visual depiction of what is being written about

Fig. 2 shows a career progression in a Fintech company starting from "Intern Backend" to "Executive/Director" with various stages in between. The path identified in this case is as follows: "Intern Backend" → "Semi-senior Backend" → "Senior Backend" → "Senior Management" → "Executive/Directors."

Notably, the flow exhibits discontinuities at the "Intern Backend" → "Semi-senior Backend" and "Senior Backend" → "Senior Management" transitions. These discontinuities suggest potential challenges in filling these positions internally.

A visual depiction of what is being written about

Another exciting discovery is the specific roles for which the company actively seeks external candidates. Fig. 3 illustrates the flow of newly hired employees recruited externally for various positions.

This insightful information reveals how the company strategically recruits new talent by relying on external sources to fulfill workforce requirements. Figure 3 clearly represents how individuals outside the organization are appointed to critical positions, showcasing the company's proactive approach to seeking and incorporating external professionals into their workforce.

A visual depiction of what is being written about

Similarly, we can identify those positions that are more unstable, where the company is experiencing talent attrition. Fig. 4 illustrates the flow of employees who are leaving the company.

This analysis allows us to pinpoint positions with higher turnover rates, providing crucial insights into areas where the company may need to focus its efforts on talent retention and engagement. By visualizing the movement of departing employees, we can better understand the patterns and reasons behind the talent losses, enabling the company to proactively address any underlying issues and create a more stable and fulfilling work environment.

Conclusion

To sum up, we've created a method that uses raw public data and a directed graph to determine career paths in a company. We’ve developed different normalization processes for job titles and seniority levels using multi-stage pipelines to ensure accuracy. This automated process allows for a quick and thorough analysis of a company. The results demonstrate the valuable insights gained from this analysis, giving a better understanding of how a company operates.

Article uploaded on 26/07/2023