Data scientists are becoming important in HR technology.
I had thought the big thing about SimplyHired was that they were good at scouring the web for job openings and putting them in one place. However, to a large extent it competes on the quality of the search engine they offer job seekers.
The quality of that search depends on their data scientists figuring out the intent of someone’s search.
Here’s an example shared by SimplyHired’s Susan Martindill:
If a job seeker only includes the keyword “director,” a typical search engine could return a variety of different results (e.g., film director, IT director, etc.). However, by examining a job seeker’s search history (e.g., did this seeker include related keywords in previous searches and/or include a location), our data scientists can create distinct algorithms that are more likely to show “film director” positions to candidates who look at entertainment-related jobs in Los Angeles, and “IT director” jobs to candidates to also searched for programming languages in San Francisco.”
If their data scientists do a good job, the search works and everyone walks away happy, with only insiders appreciating the sophistication behind the scenes.
Another company where data scientists loom large is EMSI (a CareerBuilder company). EMSI reports on labor market data; but to make sense of that data one needs a deep understanding of it. For example: Understanding why job postings don’t tell you much about the market for welding jobs.
What happens if you don’t have a deep understanding? The Canadian government got data from another vendor (not EMSI) and concluded the job vacancy rate was so low, just 1.5 percent, that Canada needed to bring in foreign workers.
Fortunately some data-savvy civil servants in the Parliamentary Budget Office discovered the growth in job postings was almost entirely due to a rise in postings on Kijiji (where multiple postings of the same job inflated the numbers). The real job vacancy rate was not 1.5 percent but 4 percent. (see The (Toronto) Globe and Mail article).
We are moving from an era where HR Tech focused on making processes better to an era where HR Tech focus on the power of analytics to help us understand the world, whether that be to understand the intent of a query, the vacancies in the job market, or something else.
What is interesting?
- Data scientists are just as important as programmers in some HR tech applications.
What is really important?
- Vendors are getting far ahead of the typical HR person’s ability to understand what they are doing. In a case like SimplyHired, this may not matter so much; all HR needs to know is that whatever they are doing it seems to work.
In cases where big decisions are being made, as with labor market data, we need to have a sufficiently deep insight into the data that we can draw the right conclusions rather than presume a number is “correct.”
This places new demands on how much analytics depth HR needs to have access to.