Dover is the first platform for automated recruiting ops, helping teams hit hiring benchmarks by identifying ideal candidates and scheduling interviews at a fraction of the usual cost.
We sat down with Anvisha Pai, Co-Founder and Chief Product Officer of Dover, to learn how their automated calibration process dramatically streamlines and strengthens recruiting workflows. We dive into:
- How to personalize outbound email campaigns while using automation
- Tips for boosting candidate response rates to recruitment comms
- Why the Dover platform is built to improve with each new client
“Through automation, Dover is able to really streamline the entire hiring process: from sourcing to candidate outreach to scheduling interviews.”
Leveraging machine learning to personalize hiring pools at scale
Every new Dover customer is taken through an automated calibration process. Steps include:
1. Establishing a baseline
Dover asks a series of business questions about the types of recruits they’re looking for – whether it’s a Marketing Manager or Staff ML Engineer. The hiring manager or recruiter answers basic questions like years of experience and skillsets, and more subjective criteria like how much they value startup experience or company prestige.
2. Integrating feedback loops
Dover then shows a series of real candidate profiles based on the outlined criteria.
The system already understands most roles across the tech industry, but this questioning process and its resultant feedback loops further calibrate to the user’s specific needs.
3. Start recruiting
Overall, this process is an automated combination of rule-based systems and machine learning, resulting in hyper-personalized results with lists of ideal candidates.
An engine growing more accurate by the day
As a system built around ML and compounding data, every new customer grows Dover’s knowledge on what comprises the ideal hire for a certain role.
In other words: After hundreds of Dover users have leveraged the platform to hire full-stack SWEs, each new search for a full-stack SWE grows even faster and more accurate.
However, the company wasn’t always automation-first, and, when they first launched, Anvisha manually handled candidate sourcing and even phone calls.
In the three years since, those tasks have been seamlessly productized.
Moving beyond the top of the funnel
After launching their flagship Outbound product, Dover saw massive growth and started to build product to manage referrals and job boards, in addition to helping their customers conduct interviews and even schedule Onsites. Dover aspires to be an end-to-end suite where customers can manage and delegate out their recruiting tasks.
Why feedback loops are essential to personalizing customer experiences
The Dover product relies on both personalized candidate criteria and constant scaling to serve the labor markets and their ever-growing client base.
More specifically, that personalization comprises two elements:
- Transparency for users — Dover avoids a black-box approach. There’s reasoning behind every candidate shown to every customer; they can peek under that hood whenever.
- Continual feedback — Like any intelligent system, Dover is constantly learning, collecting feedback and adjusting parameters accordingly each time a user rejects a candidate.
Again, before Dover’s tech had progressed, their customer success team would manually check in with clients and request calibration feedback.
“Dover is unique because we’re not just a machine-learning black box doing random tasks. Our system has a progressive understanding of what various roles entail.”
How Dover streamlines recruitment: From listings to candidate comms
Once Dover has been calibrated to a customer’s set of needs, their hiring pipeline is essentially entirely automated, managing both inbound and outbound applicant traffic.
- Outbound — Clients specify ideal candidate criteria. Then, Dover reaches out on their behalf and builds an automated recruitment flow (similar to sales automation). It can also suggest ideal matches within the team’s networks.
- Inbound — Dover automatically posts listings to external sites and flags top inbound applications.
The platform will regularly make recommendations on this front, such as:
- Suggesting compatible job boards for the listing (i.e., Women Who Code for engineering)
- Auto-generating outreach copy or job descriptions using GPT-3
- Enabling A/B testing (built into the platform) for trying out varied recruitment methods
Tapping existing networks for high-quality outbounds & referrals
An inbound applicant who’s made contact via a job board is clearly already interested in the role.
Meanwhile, for warm leads, Dover runs through a separate automated recruitment process:
- Everyone at the client company uploads their connections to Dover
- Dovers combs the data, identifying ideal candidates from the smaller pool
- The client’s hiring manager can simply click a button to initiate outreach
In Anvisha’s experience, referrals via personal connections yield higher response rates but less conversions than inbound recruits with demonstrated interest — mainly since the candidates are more passive.
Integrating deep learning into candidate comms
One major use case for the API has been classifying inbound replies from candidates, such as:
- “Sorry for the late reply; I’ve been out sick”
- “No, thank you. I’m not interested.”
- “I’d need a visa to take this role.”
By classifying responses on a linguistic-emotional spectrum (from interested to not interested), Dover customers can more rapidly draft replies to candidate comms. We track this metric directly in-app, so customers know what their “interest” rate is for any open role.
How to optimize candidate engagement in email campaigns
As for tactical ways to drive recruitment email success, Anvisha recommends the following:
- Capture the company — Is the recruit compelled by your company’s mission or product? Detail these elements through social proof, revenue or impact numbers, etc.
- Impact on their career — Would this role be a promotion or a lateral move within the space they’re interested in? Could this role boost their career trajectory?
- Personalization — Of course, the final push for a candidate to accept an offer requires feeling uniquely seen and valued by their potential employer.
In Dover’s case, they’ll use intelligent text content based on the candidate’s resumé to personalize emails (i.e., by including unique references to their past experiences).
Customers are shown to receive stronger response rates in doing so.
Additional methods for personalization, naturally, include Windsor. Imagine a candidate opening up their offer email to a video from the company’s CEO, welcoming them by name to the family.
Anvisha highlights some major recruitment insights they’ve accumulated over time:
- Directly address the recruit — Including the candidate’s name in the subject line significantly boosts open rates and, as such, has become common practice.
- Focus on the work — Discussing the actual role your company is recruiting for (as opposed to the company itself or its investor pool) drives a much higher response rate.
- Up your email game — Make info in emails easily accessible with bullet points; always link out to press and social proof; and ensure emails are mobile-friendly.
The Dover playbook: dos and don'ts for driving reply rates
As a platform, Dover frequently encourages customer experimentation. Over time, they’ve accumulated strategies proven to improve response rates. A couple include:
- Recruiting — Recruiting a candidate from an underrepresented group in the industry yields greater success when the recruiter shares that demographic identity.
- Compensation — Aside from transparency on salary, people want to know if the role is remote or perhaps based in a city they’d want to relocate to (with the help of a stipend).
On the other hand, with each new client, the Dover team sees recurring blockers that hinder success with recruitment campaigns. Anvisha outlines a few below:
- Too much text — No one needs an encyclopedic job description in the first email.
- Zero credibility — If you’re the CEO of a smaller company emailing a top candidate, don’t forget to identify yourself to drive home your interest and drum up excitement.
- Generic subject lines — Make your subject line unique (again: start by including their name), since qualified candidates are likely already getting tons of recruiter inbounds.
- Excluding key info — Questions they’ll need answers to include: Is the job remote? What’s the salary? What coding languages are required?
“When it comes to recruiting, personalization at scale is everything. I’m excited to see technologies like Windsor that unlock new levels of personalization.”