Building a data-driven hiring machine | by anton_g

Attracting talent at scale is more than just data. But a data-driven process makes it a lot easier


In this article, I would like to pass along some of my learnings in setting up a hiring machine for scaling an (engineering) organisation. In fact, these learnings are applicable to any hiring process you would need to setup at scale. There are many ways to build a recruitment machine and I am sure that what you will read in this article is not exhaustive or either the best possible way.

This article could be most interesting to you if you are in the following state:

  • An organization of 20–50 people ready to scale that has some processes in place but not enough capacity (recruitment power)
  • Pressing needs to hire on time to support the business plan
  • Need for hiring senior engineers and talent that can level up the overall expertise in the organisation or bring skills that are not yet well-developed in house
  • Lack of a (structured and) data-driven hiring process
  • A very competitive hiring environment — not enough available people that you might consider a fit and many companies that are targeting them

A lot can be written about recruitment and hiring at scale. Here I will cover the following topics.

  • Treat the hiring pipeline as a sales pipeline
  • Be candidate-centric in the same way as you are customer-centric
  • Attract talent to retain talent
  • End-to-end automation
  • Continuous training and learning for interviewers

Many will start with other key aspects of recruitment but I would like to draw your attention to the data you need in order to have highly efficient and successful hiring.

Why not start with any of the rest?

One has to measure to know. If the “temperature” of the hiring pipelines is not continuously measured, there is most likely lots of time wasted and the recruitment team is running blind and inefficient.

When hiring 3–4–5+ candidates per month/week in an environment where you do not have enough recruiters, sourcing capacity, budget like the Big Tech companies, and engineers to conduct interviews, you have to be extremely efficient and use data to understand as early as possible if you are on track. This will allow you to apply corrective actions timely and keep everyone on the same page. The same way you will do with a sales pipeline that drives revenue into the company. When hiring at scale the success of a company depends on the team’s ability to hire on time the roles most needed. With that said, please do not compromise with the amount of data and analysis you use.

What are some useful metrics that can help build a data-driven pipeline?

Conversion rate per stage. It helps to determine the health of the hiring pipeline. Change in the conversion per stage can lead to less/more people hired and often can indicate a change in the environment or some inefficiency in the process. Such changes help to identify bottlenecks in the pipeline and the team can experiment to improve the hiring process. It also allows the hiring team to balance their effort between the different pipelines.

  • Conversion rate for different job descriptions (JD). Different JDs for the same role can have different performances when attracting talent. It is not easy to measure but you can experiment with 2–3 JDs for the same role on the same channel and see which one performs better (bring more quality candidates to the pipeline). You can also decide using a slightly different JD per channel if that makes sense but this requires a little more effort than the hiring team might not have enough capacity for. This can be further developed and such experiments can be performed for the content of the sourcing messages, the message mediums (text, video), and the person approaching candidates.

Time to hire. This is usually a very important metric when a company wants to hire according to their business plan. It can be used as an objective of X days to hire a candidate per pipeline. An interesting way to actually control the speed of hiring is to measure Time in stage.

  • Time in stage. The time from a first conversation to an accepted offer can take weeks. A way to measure and improve the speed of the hiring pipeline is to set a maximum time for a candidate to be in a specific stage and an average time across all stages. This is measured for all stages except sourcing contact and offer stages. This metric usually shows the “bottleneck” stages in the pipeline. Analyzing why a stage takes longer than desired and applying corrective actions helps to speed up the pipeline and shorten the Time to Hire.

% of new hires above the average. This is a quality metric that can be used to measure the quality of the candidates compared to the engineers in the team. This metric requires an organisation to have a well-defined professional growth system that is actively used to assess candidates during the hiring process. Such a system will allow the hiring team to know the level and expertise of the different people at the different levels in the organisation and help to determine if a candidate is better on average or if they bring anything that the team has less of.

Hiring Rate (conversion rate). This is the conversion rate of the hiring pipeline. If 100 people entered the pipeline and 3 were hired, the Hiring Rage is 3%.

Hiring Rate per acquisition channel. Most companies use different sources to bring in candidates. These channels will attract candidates that fit your needs or do not. As those channels are often very expensive (subscription fees and more), it is key to understand the value each channel adds to your hiring effort. Once, you identify the best channels you can manage your acquisition channel portfolio and increase the overall hiring rate.

Hires per week per role. When hiring at scale to support a business plan, there will be a number of people in different roles that are needed to start in a specific month. This is very hard to time well but measuring hires per week allows the team to balance their hiring effort, keep a good pace and have a better forecast for future hiring plans.

Investment to hire. Cash is never unlimited and hence, understanding the cost invested to secure a new hire is important. Even if this is not required by the finance team or management, it is important to keep an eye on this metric. Tracking per pipeline is interesting to understand the investment needed to make a hire for each specific role.

The aforementioned metrics are not exhaustive and depending on the situation more metrics can be implemented. For instance, if there is a significant sourcing effort by the team, it is important to measure candidates’ response rate (first call), and more.

In general, these metrics should be applied across all pipelines (at least in the same organization) and can be used to balance your recruitment, learn and optimize on the fly.

You can also perform experiments to see what works and what doesn’t. Those can be done similarly to a/b testing but usually, it will be very difficult to have conclusive results as most likely the number of candidates in the pipeline will not be in the number of 10s of thousands. Still, it can give you some good indication.

1*nlMawG 6G7RjCLIkL2X40g

There is a lot written and discussed being customer-centric. Many companies are doing a lot to help their teams focus on the customer. If it is a customer we will hand-hold them throughout the onboarding of our product and do everything we can to make them successful. We will provide them with information and with any support they need. We will sell them the opportunity and articulate the business case — the value our product will deliver to them. IMO, the same should be the approach with candidates — we should treat them as if they were our customers.

Let me share a few of the things that I found really working well and that do not require that much effort to put in place.

  • Hiring managers should be involved end-to-end in the process and speak as early as possible to a candidate. They know the role best. They know who their team needs. They can best connect with a candidate and help them learn about the team and see if the candidate is a fit. They are probably the best sales manager in a company for a specific role; that is the product we are selling after all.
  • Document your process including timelines and made it public. It is good for a candidate to understand what to expect and it is good for your team to make a public commitment to the time it takes to go through the process including each step.
  • Document each step in the hiring process outlining what the candidate should expect. From my experience and feedback from many candidates, this is really helpful. Candidates know what exactly the interview will be and what to prepare. It helps them prepare and be at their best during the interviews. I consider this very important because time is scarce and missing good candidates should not be an option.
  • Publish your compensation model (if you can). People appreciate transparency and they know that they will be paid fairly based on their skills and experience. In my experience, this makes the conversation about compensation very constructive and much easier.
  • Publish a few articles that will help a candidate to better know your company and teams. It could take a few weeks to produce such articles but this is really useful. It triggers additional questions and allows a candidate to ask something else than the standard “How does team X work?”.
  • Have a talent acquisition specialist follow up with all candidates throughout the process providing any additional information and helping the candidate to make an educated decision. Making a personal connection with each candidate is important. It helps a candidate to be more successful and improves the odds of hiring them. It requires some effort and hence automating the hiring process as much as possible is important.
  • Before extending an offer, use the connection you have built with a candidate to pre-discus it and make sure that all works for them as well as for you.
  • After extending an offer make sure to stay connected. It is not a done deal even when the offer is accepted. This is a great opportunity to engage with future employees and connect them with colleagues and/or provide them support if they will be relocating. A couple of follow-up calls work great. Sending a welcome package is great too. Letting them talk to a couple of people from your team is helpful and helps you differentiate.
1*eYoeeDaq rHbcwE75NNg4g

It is well-known that engineers like working with engineers they can learn from and whom they can be challenged. Many of us have heard engineers sharing that they have no one to challenge their views or no one to learn from. This is because in our domain things advance very fast and we have to often deal with complex problems to which many solutions are applicable. Having people to bounce ideas with and learn from makes work much more fun and retains talent.

When scaling very fast often recruitment teams compromise on the quality of the candidates or their cultural fit in order to hit the hiring targets. Looking for and finding the right people can slow down the process by increasing the Time to Hire but “wrong” hires are much more costly.

There is a couple of components that can be added to the hiring process that helps reduce the chance of hiring people that are not a fit or not at the required level.

  • Define a % of all hired candidates that you want to be above the average for a function and/or bring something that you have to the team. This is a metric that is easy to implement if you have a good and detailed growth system.
  • Concrete aspects of the role (skills, experience) to be covered at each interview by each interviewer and detailed interview notes with a conclusion. These could be predefined questions to ask to verify certain knowledge and skills. Focus on the experience that you really need for the role and avoid having unstructured interviews and asking questions such as “Can you walk me through your CV?” or “Can you tell me something that is not on your CV?”. Honestly, those add zero value.
  • Make sure each interviewer has detailed notes from the interview and has their conclusion in written form in the ATS. Data is key and having detailed notes helps to evaluate the candidates’ experience, skills, and fit with the team.
  • An “independent” interviewer. Have a senior interviewer (e.g. another manager) that is not part of the team the hire is for. This person should be the one who has a veto and should determine a good fit for the role. I “stole” this idea from Amazon a few years ago and IMO this really works well. Amazon calls it a “bar raiser”. You will need a few such experienced interviewers in your organization that can handle the role and help you train and coach the interviewing team.
1*vG 5n5kDg yPwLTtwg3oHQ

When an organization is starting to scale and doesn’t have enough experienced interviewers it is important to set up a training program for technical interviewers. The more the organization grows the more interviewers will be needed and available. Hence, having a unified approach to interviewing and a continuous training program helps the organisation scale. Soon new joiners will be interviewers and they will be hiring people they think are a fit. If you would like to maintain the same quality of hiring and hiring for the team that you have been building, continuous training and learning is important. Here are a couple of suggestions that I have seen working.

  • Educate your hiring managers to define an interview process that clearly outlines the role, the ideal profile, what skills/expertise/ core value each interviewer will be checking, and the interview approach (open questions / behavioural questions / etc.).
  • Educate the hiring team on how to avoid bias. It is important to hire a diverse team and it is well-known that people tend to hire people like themselves. Structured interviews and detailed notes help to reduce the bias a little as everyone can read the notes and make their own conclusion. Making sure to have a diverse hiring team helps as well. If different people interview a candidate throughout the process the biases each interviewer has are presumably less affecting the process.
  • Practice continuous feedback with the recruiter. Every new role will take some time for the talent acquisition specialists to adapt to and for the hiring manager to polish the requirements out. Hence, having a practice of continuous feedback on candidates that the hiring manager has interviewed is key for the success of the hiring effort.
  • Learn by interviewing. Set up a program for engineers to become technical interviewers. Shadowing works really well. It is a concept where an experienced interviewer conducts an interview and a new interviewer is shadowing to learn. It can be two shadowing sessions for a new interviewer an experienced colleague and two sessions in which the new interviewer conducts an interview and the experienced one shadows. After each interview, it is important to debrief and share feedback and learnings. After that every month for three months the new interviewer is shadowed and feedback is shared. Note that not all people going through the program would like to be interviewers or should be accepted as interviewers.
  • Feedback throughout the interview process. Let the hiring managers to share feedback on the detailed notes of other interviewers and on their conclusions. Such feedback is very valuable as it allows the interviewers to learn from their own work and to better understand the requirements for the role.

Every minute spend on mundane work is a minute that can be used for connecting with and sourcing talent. Many organisations use some rudimentary tools or even use google sheets (yup, it was a surprise for me as well) to manage their pipelines. Imagine a situation where you use one tool to manage coding challenges for the candidates, a sourcing tool, an ATS, and a reporting tool of your taste. Only mapping the results from one system (coding challenge or sourcing) to the ATS will cost up to a minute of the recruiter’s time excluding any distractions. Now imagine you have an inflow of 100s of candidates weekly per pipeline…

Most companies hiring at scale will most likely need tools for sourcing, coding exercises, ATS, and more. Most tools offer out-of-the-box integrations but often that is not enough. My experience with various tools has led me to believe that metrics and reporting are often a second- (maybe even third-) class citizen of those systems.

In general, my recommendation here is based on the fact that I haven’t seen any one tool that does everything and really has metrics and reporting tooling as a first-class citizen.

Choose tools that offer extensive APIs to extract data (for reporting), to manage any of the candidates and pipeline data, and offer integration via webhooks or other mechanisms to notify for changes such as a complete coding challenge, or change of a candidate’s hiring pipeline state.

Evaluate the level of automation a tool offers.

  • For ATS pipeline automation such as notifications, automatic interview scheduling, tagging of people, automatic emails to candidates, notification delays, etc.
  • For sourcing tools, it is important to have nice and configurable options for the messages sent to candidates allowing different cadence and templates.

Once all of this is in place integrations and automation can be implemented at any time by observing the bottlenecks and the steps that are still manual and take time.

Most importantly define the metrics that will guide the hiring process and configure reports asap. The earlier those are in place, the earlier the learnings will come and the faster the hiring team will adapt.

This article is by no means exhaustive and only aims to share my experience with hiring in recent years. The idea is to encourage the readers’ thinking in the direction of data-driven hiring with fully integrated and automated systems. IMO, this is the only way to hire at scale.

Honestly, I have never been able to completely crack the nut of hiring even when using data. Sometimes there are changes in the performance of a pipeline without any obvious reason that comes from looking into the data. Attracting talent at scale is more than just data but without a data-driven approach, I believe, it is absolutely hard to achieve.

News Credit

%d bloggers like this: