Abstract

Technology trends play an important role in the hiring process for software developers. Recently, various sources point out misunderstandings and bad practices in this area. The anecdotal term Résumé-Driven development (RDD) describes a phenomenon of overemphasizing technology trends in the application process. A current study conducted by the University of Stuttgart examined this phenomenon empirically by interviewing 591 software professionals, in both hiring (130) and technical (558) roles. The results reveal RDD facets in essential parts of the sample: 60% of the hiring professionals agreed that technology trends influence their job advertisements. Among the software professionals, 82% believed that using trending technologies in their daily work would make them more attractive to potential future employers. The article highlights the results of this study and discusses influencing factors as well as consequences for the practice of software development in general.

Introduction

"Specific technologies over working solutions, hiring buzzwords over proven track records, creative job titles over technical experience, and reacting to trends over more pragmatic options" — this vision of the future of software development is drawn from a satirical manifesto about Résumé-Driven Development[1].

In fact, there are increasing signs that this admittedly gloomy vision of the future can already be found in practice today. In times of social networks, communities, job portals and especially career platforms, the own profile including a professional resume becomes more of a figurehead than ever. A decade ago, such personal details were considered confidential information, disclosed only in the process of applying for a new job position. An up-to-date profile on platforms like Xing or LinkedIn comprising professional career, degrees and certificates obtained as well as knowledge and skills confirmed by colleagues (“endorsements”) It is nowadays rather the rule than the exception and completes a professional appearance.Moreover, a high number of personal contacts increases the profile's credibility and exposes potential impostors, likewise.

Thanks to sophisticated search capabilities, those who present themselves appropriately and comprehensively on such platforms will be found first — a real treasure trove for recruiters and companies, as many of us can certainly confirm. What at first glance looks like an extremely useful tool for applicants and hiring professionals has its downsides as well.

Apart from general concerns about digital abstinence or data protection, the great importance of profiles and resumes being constantly available to almost everyone also leads to an increasing urge to perfect one's own appearance. For software developers, knowledge and skills regarding various technologies play an important role here. A recent survey among 65,000 software developers showed that technological skills are seen as the most important factor in the application process [2]. At the same time, a study using eye tracking found that hiring professionals spend an average of 7 seconds looking at an applicant's resume [3]. This leaves little time to convince the recruiter's critical eyes. It stands to reason that this is most likely to succeed with breadth and topicality of the technologies presented (cynical contemporaries may speak of buzzwords here). Hence, a wide-ranging repertoire of technologies increases the likelihood of being found, e.g. through search queries. 

The phenomenon of Résumé-Driven Development

A strategy to build an impressive personal profile seems clear: The selection of technologies is given highest priority in the current field of activity. Moreover, future employers and jobs may specifically be chosen according to these criteria. Frequent changes are inherent to not get stuck in one topic. Unfortunately, the actual requirements in the current development project take a back seat then. A satirical term for this phenomenon has recently emerged in developer circles: Résumé-Driven Development (RDD). Its essence lies in a focus on current technology trends that fill gaps in the applicant's profile, thereby completing it and making it appear more impressive. Focusing on the profile's attractiveness inevitably supersedes or replaces project-specific requirements which should in the first place determine the technologies to be used.

If I as a developer lack practice in frontend development using Angular or React, I will prioritize one of these frameworks in the current project. If there is still too little about microservices in my resume, perhaps I will implement this simple web app usingSpring Boot and deploy it as container solution in a Kubernetes cluster. The preference here is based on current trends or hypes which not seldom disappear from the market after a few years. In addition to media, developer communities and forums, the regularly appearing Gartner Hype Cycle [4] is a good indicator here. It allocates each technology trend to one of five phases that it typically passes through. An even more precise seismograph for such technology movements is the quarterly technology radar issued by ThoughtWorks [5].

However, developers in their role of applicants represent only one side of RDD. On the other side we find companies represented by their HR departments who act or react accordingly. If a developer is focused on his/her own profile while the company aims to create the best possible product with the help of the right tool, then these diverging interests can easily lead to conflicts in the choice of technology. Companies are indeed on the horns of a dilemma here. The technologies that they search after may often be the ones that developers are NOT particularly keen to work with. Since good software developers are rather rare on the job market [6], one may be willing to make concessions. RDD on the side of companies and hiring professionals now implies that they upgrade their job postings using current technology trends that are popular with developers. That way, they might address a larger number of applicants and perhaps find a better suited candidate in the end. In return, applicants will ultimately assume that the companies' job advertisements reflect the actual needs of an employer and hence feel encouraged to optimize their profiles even more with regard to a wide range of current technology trends.

As we can see,Résumé-Driven Development describes an interaction between the applicant and hiring sides with a focus on technology trends. It is very likely that such behaviors will eventually have negative side effects. On the one hand, it can affect applicants whose expectations are not met (see chapter "Why SoftwareDevelopment Hiring Is Broken" in [7]). The long-term consequences for companies, on the other hand, can be even more serious, if poor quality and maintainability of the created software result in high costs.

The legitimate question arises whether these interconnections outlined above actually exist and whether causality can be proven. It would also be interesting to appraise the extent of this phenomenon in its current presence and the expected effects on software development in practice. 

A scientific study

There has been no scientific research on this subject done so far. Moreover, there is no clear definition for the term Résumé-Driven Development (or CV-Driven Development).It can be sporadically found in blogs and discussion forums on this matter, where it often leads to controversial debates or even polemics [8, 9, 10]. A research team from the Institute of Software Engineering at the University ofStuttgart in Germany has set itself the goal of bringing light into the darkness. An empirical study launched in 2020 aimed to explain the phenomenon with scientific methods and to determine its influencing factors. Another goal was to introduce the notion of Résumé-Driven Development in the scientific discourse. A wide-ranging survey was started among software developers on one and hiring professionals on the other side. Developers in their various roles, including students from computer science degree programs have been consolidated to the applicant group. The hiring group consisted of employees from HR departments, managers with personnel responsibilities and specialized headhunters in the IT industry. Participants were able to respond in the combined survey even from two perspectives, if applicable.

Surveying these two groups shaped the foundation of the study design through investigating the recruitment process, with a focus on the candidates' CVs and hiring's job advertisements. Based on a sufficiently large sample, specific intentions and motivations of both sides should be yielded, along with further influencing factors on the phenomenon. The web-based survey was online from May to July2020 and achieved a trans-regional spread, not the least thanks to the support of a leading German developer magazine. Through other developer forums, career networks and last not least industrial contacts of the researchers the spread was even broadened internationally. In the end, a total of 591 participants could be counted, with the majority being located in Germany. Divided into the two perspectives of hiring and applicant, the answers resulted in a ratio of130 to 558. Demographic data showed a realistic distribution in terms of professional experience and company size: beginners with less than 2 years of experience and experts with more than 20 years each made up one fifth, with the remainder lying in between. The dominant job role was software engineer, followed by manager, student and architect (see Table 1). The industry affiliation was dominated by the Software & IT Services domain, followed by Automotive. 

Results for the hiring perspective

The hiring participants were first asked regarding their preferences on knowledge and skills of applicants. The first queried aspect was the preferred technology-related orientation “breadth” vs. “depth”. Around two thirds of the respondents considered both characteristics important. However, given a choice,22% of the participants would prefer a candidate with in-depth knowledge in specific technologies while to 42% the candidate with broad knowledge in a variety of technologies appears more attractive (see Figure 1). This shows a slight tendency towards a generalist profile. Around one third of the respondents could not make a general statement on this matter or had no preference.

The second aspect targeted the technology-related characteristics “established” versus "latest/ trending". 59% of the participant preferred the latter, while established technologies were more important to 85%. Once again, put in direct comparison, a significant advantage for knowledge in established technologies(39%) over knowledge in current technology trends (20%) showed up. Here as well,41% could or did not want to make a general statement.

 We can conclude that broad knowledge in established technologies tends to more likely meet the needs of hiring professionals and thus companies. Another important aspect in the interaction of both groups constitute the expectations from / assessment of the other side(see Figure 2). With regard to the applicants, 71% of the hiring participants agreed that software developers generally enjoy working with the latest /trending technologies. In addition, about half of the respondents assumed that applicants would rather be discouraged by the prospect of working with established (“legacy”) technologies. Interestingly, a majority (59%) of the hiring respondents also agreed that technology trends have an impact on the contents of their job advertisements. Specifically asked if their advertised technologies are influenced by the expectations of potential applicants, was still admitted by a large fraction of 46%.

 These results show obvious indications for the presence of RDD. More difficult was the identification of concrete reasons and causes. One obvious influencing factor might be the labor market: 59% of the hiring respondents experienced difficulties in finding suitable candidates for their vacancies, while only 15% indicated that it wasn't a problem for them. It should be expected that companies compete for the best applicants on the job market. That this fact actually causes the companies to make the above described concessions, could however, not be proven by a linear regression analysis of the survey data.

Results for the applicant perspective

The applicants were first asked regarding the role that technology trends play for them in general and second while creating their profile or CV (see Figure 3). A large majority of 73% stated that they enjoy using latest and trending technologies in their daily work. This confirms the perception uttered by the hiring side, with an almost equal percentage. Only 18% found it challenging or uncomfortable to keep themselves up to date in this regard. However, only 42% of the respondents were convinced that they would become better software developers by using such technologies. In addition to this intrinsic motivation for using trending technologies, 82% of the applicants were convinced that such knowledge and skills would make them more attractive for potential employers. Moreover, 63%confirmed that an even higher variety of technologies would further increase their own attractiveness.

Subsequently, the applicants were asked regarding their practical experiences with technology trends (see Figure 4). Half of them reported about predominantly positive experiences. For 19%, these experiences were rather negative, 32% responded neutral. Finally, one fifth of the respondents admitted that they have already used a trend technology, although it was not the most appropriate choice for the concrete project or application.

In summary, these results suggest the existence of RDD also on the applicant side: Technology trends do not always prove beneficial in practice, but are considered significantly more important when increasing the own attractiveness in the hiring process.

A definition of Résumé-Driven Development

The exploratory study also aimed to propose a definition of the term Résumé-Driven Development in the scientific discourse. Figure 5 shows the developed construct for this purpose. On the left side, we find both interacting groups, named “Hiring Perspective” and “ApplicantPerspective”. The following two facets characterize hiring: The degree to which1) technology trends and 2) the expectations of applicants influence their job offerings. Whereas, the presence of the applicant perspective was shaped by 1) the degree to which applicants are convinced that knowledge of trending technologies make them more attractive for companies and 2) the importance of trends / hypes in the choice of technology. The thereby shaped RDD construct is reinforced by the strengthening predictors on the right, which were determined through a regression analysis.

The combination of the facets in both groups — and thus Résumé-Driven Development — is reflected by a substantial part of the sample in the interview study. An important aspect here is the fact that hiring professionals actually tend to prefer knowledge in established technologies. Further analysis found a positive correlation with the company size. However, the data did not show any significant correlation with the labor market situation (based on the information provided by hiring participants about their perceived difficulty in finding appropriate candidates).

Consequences and outlook

The question arises, what potential effects Résumé-Driven Development has on the practice of software development. Here we can distinguish two main areas. First, long-term effects on software quality are very likely. The majority of participants on the applicant side agreed that constantly emerging technology trends increase the diversity of languages, frameworks and tools used in the company. This boosts the complexity [11] and hence affects the maintainability of the developed software, either through the management of dependencies and updates or the necessary knowledge transfer within the team. In addition, a lack of reliability is often attributed to immature technologies. Since emerging trends do not seldom disappear after a few years, the effort can quickly turn out as a bad investment. For companies, RDD can therefore have a negative impact on quality and reliability and moreover drive up maintenance costs. The consequences are not immediately apparent, but manifest in the mid- to long-term. A recent article in the IEEE Software Journal on the subject of "The future of software development" points to already visible consequences of "overwhelming complexity, combined within sufficient development competences", for example in "public's decreasing acceptance of […] self-driving cars" [11].

The second area of impact concerns the recruiting process itself. Here, RDD can arouse false expectations on both sides. Applicants generally do not like it if their future role in the company is not clearly defined [12]. Inadequately communicated criteria by hiring were identified as one of the main deficiencies in the analysis of over 10,000 reviews on the Glassdoor career portal [13]. A subsequent high fluctuation is costly for both sides, but especially for the company: The cost-intensive training was a false investment and in the worst case the newcomer leaves behind code in a cutting-edge technology that none of the other team members is (or will ever be) able to maintain.

Another point that was mentioned often in the free text answers to the Stuttgart study is the associated neglect of soft skills. These include social skills such as communication, self-motivation and the ability to learn, as well as a basic understanding of the principles behind the various technologies. It also reflects the observations made by the author already quoted above who knows both perspectives from his personal experience [7].

There are always two sides of a coin. Hence, it is of course possible to derive positive aspects from this phenomenon. The history of software development was always characterized by permanent change and the ever faster expanding technology landscape requires a constantly high level of willingness to learn. Developers who follow technology trends and keep themselves up to date, possess such skills and prove their ambitions thereby. For a company, it may be more lucrative in the long term to give preference to an applicant who is particularly willing to learn, even if he/she does not perfectly meet the requirements in terms of his/her current knowledge. Besides, the inquisitive newcomer can serve as a source of inspiration for experienced team members and breathe new life into an established team.

Finally, the renowned computer scientist Donald Knuth shall be quoted: “computer programming is an art […] especially because it produces objects of beauty. A programmer who subconsciously views himself as an artist will enjoy what he does and will doit better ”[14]. In a way, it speaks in favor of giving software developers the freedom to choose their preferred tools and technologies. However, today's omnipresence of algorithms and software in many safety-critical systems requires strict quality standards and engineering methods. While this shouldn't discourage anyone from following trends and regularly acquiring new knowledge, the best places to train and demonstrate such skills are for example coding challenges or hackathons.

For further details on the subject, we refer to the full article "Résumé-Driven Development: A Definition and Empirical Characterization" that was presented at this year's International Conference on Software Engineering (ICSE 2021) [15].

References:

1 "Resume Driven Development", 2020.  Available: https://rdd.io

2 Stackoverflow, “2020 Developer Survey”, 2020.  Available: https://insights.stackoverflow.com/survey/2020#

3 Ladders, "Eye-Tracking Study", 2018.  Available: https://www.theladders.com/static/images/basicSite/pdfs/TheLadders-EyeTracking-StudyC2.pdf

4 Gartner Group, "Hype Cycle", 2021.  Available: https://www.gartner.com/en/information-technology/glossary/hype-cycle

5 Thoughtworks, "Technology Radar", 2021.  Available: https://www.thoughtworks.com/de/radar

6 Bitkom, "Erstmals mehr als 100.000 unbesetzte Stellen für  IT-Experten", 2019 https://www.bitkom.org/Presse/Presseinformation/Erstmals-mehr-als-100000-unbesetzte-Stellen-fuer-IT-Experten

7 W. Gant, "Surviving the Whiteboard  Interview", Apress, 2019, ISBN: 978-1484250068

8 M. Loukides, “Resume Driven Development - O’Reilly Radar”, 2014. Available: http://radar.oreilly.com/2014/10/resume-driven-development.html  

9 S. Teller, “How CV-driven development shapes our  industry - A geek with a hat”, 2020. Available: https://swizec.com/blog/how-cvdriven-development-shapes-our-industry

10 S. Jha, “Why we at $FAMOUS_COMPANY Switched to  $HYPED_TECHNOLOGY”, 2020. Available: https://saagarjha.com/blog/2020/05/10/why-we-at-famous-company-switched-to-hyped-technology/

11 C. Ebert and S. Counsell, “Toward Software Technology  2050”, IEEE Software, vol. 34, no. 4, pp. 82–88, 2017. doi: https://doi.org/10.1109/MS.2017.100

12 HackerRank, “2019 HackerRank Developer Skills  Report”, 2020. Available: https://research.hackerrank.com/developer-skills/

13 M. Behroozi, S. Shirolkar, T. Barik, and C. Parnin,  “Debugging hiring: what went right and what went wrong in the technical  interview process”, in Proceedings of the ACM/IEEE 42nd International  Conference on Software Engineering: Software Engineering in Society, 2020. Available:  https://doi.org/10.1145/3377815.3381372

14 D. E. Knuth, "Computer Programming as an  Art", Commun. ACM, vol. 17, no. 12, pp. 667–673, Dec. 1974. doi: https://doi.org/10.1145/361604.361612

15 J. Fritzsch et al., "Résumé-Driven Development:  A Definition and Empirical Characterization", 2021. Available: https://doi.org/10.1109/ICSE-SEIS52602.2021.00011