Kieran Snyder

CEO at Textio

THIS CHAT HAPPENED ON December 15, 2016

Discussion

Kieran Snyder@kieransnyder · CEO, Textio
Hi! I'm Kieran Snyder, the CEO and Co-Founder of Textio, a predictive engine that tells you who's going to apply to your job posts based on the language you use. Textio finds insights in large document sets of any kind, and pinpoints the patterns in your writing that get provable results. I have a PhD in natural language processing and led product and engineering organizations at Microsoft and Amazon before founding Textio. I've written about bias in business documents in places like Fortune, Re/code, and the Washington Post. I love talking about entrepreneurship, NLP and ML, and all things talent and team-building. Excited to chat!
Philip Kuklis@philipkuklis · Co-Founder, Hubble
Hi Kieran, you’ve recently mentioned that over the past 2+ years you had zero turnover in your team (Congrats!). What makes Textio’s culture so special and how do you plan to preserve it while growing your business?
Kieran Snyder@kieransnyder · CEO, Textio
@philipkuklis Thanks, it is something we are really proud of. There are two things we look for in everyone we hire regardless of role (and as you'd expect, they're in all our job posts). We look for people with a history of trying and learning new things - because every day you're going to come to work at any startup and have to do stuff you don't know how to do. That has to be energizing rather than overwhelming. The other thing is we look for people who have a pov but are low ego. Both are really important. Best way to preserve is to maintain the hiring bar on both of these attributes!
Clare Sayas@claresayas · Tech PR @ Revere
I'm curious about your personal opinion on how businesses can make the "soft" job search more inclusive - or the practice of looking for candidates through personal connections only. Anecdotally, I have never gotten a job via cold application, but rather from people I knew from school, internships, etc. I've been very lucky, but recognize that practice can breed a monoculture within a company. How do you help break execs from that model?
Kieran Snyder@kieransnyder · CEO, Textio
@claresayas Word-of-mouth is going to be a significant factor in hiring. People like to recommend the people they've enjoyed working with before, and as a hiring manager, you should take advantage of that. People who love their jobs enough to recommend their workplaces to colleagues they value - this is awesome affirmation of your environment. Given that this is always going to be a factor, the only way to avoid a monoculture is to make sure that the people on the ground represent a variety of networks - so the colleagues who get referred come from a broad range of places. Naturally, intentional recruiting plays a big role here. If word-of-mouth is the only strategy, you're not likely to be happy with your team composition in the end.
Emily Hodgins@ems_hodge · Community and Marketing, Product Hunt
Hi Kieran thanks for joining us today. Can a workforce become more diverse through using the correct language in job postings to attract a more diverse set of candidates?
Kieran Snyder@kieransnyder · CEO, Textio
@ems_hodge The data says yes, your team can become more diverse by changing the language you use to talk about your job! Customers who get above a 90 (out of 100) in Textio on their jobs see an average of 12-15% more applicants from underrepresented groups. Not only that, but jobs where a woman is hired at the end of the process contain twice as much "feminine-tone" language as "masculine-tone language" in the initial job post. (These are phrases that statistically predict a higher % of women or men respectively applying to your job.) Makes sense - you're much more likely to hire a woman at the end of the day if you have more in your pipeline in the first place.
Ayrton De Craene@ayrton · Code @ Product Hunt
How do you begin to define a companies culture from day one? Why is this important?
Kieran Snyder@kieransnyder · CEO, Textio
@ayrton Know what you value, and hire people who exhibit those characteristics. A company's culture may be set by its leadership, but it is carried by the people who work there. And if you're successful, the founders are quickly just a tiny % of the team. Make sure the team is right, because they are the ones who really show whether the company is a good place to work.
Ben Tossell@bentossell · newCo
Do you use your knowledge of NLP on a day to day basis?
Kieran Snyder@kieransnyder · CEO, Textio
@bentossell I do. Back in college I double-majored in linguistics and math. Everything I've done since has built on that. Most of my tech career has been in NLP, internationalization, or the search space somewhere. As a CEO I spend a lot of time with our data team working through models. I designed the first predictive model that Textio used!
Ben Tossell@bentossell · newCo
If you had to swap lives with a tech CEO for a week, who would it be and why?
Kieran Snyder@kieransnyder · CEO, Textio
@bentossell I wouldn't swap lives with anyone. =) But if I could be a fly on the wall of their brains, Elon Musk or Jeff Bezos. They have clear conviction about what they are making, and have seen their companies through every stage of growth.
Jackye Clayton@jackyeclayton · Jackyeclayton@gmail.com
I see great potential for enhanced diversity recruiting capabilities. Is a more diversity focused module on your roadmap?
Kieran Snyder@kieransnyder · CEO, Textio
@jackyeclayton Great q. Textio exposes gender in our UI right now because our data set that allows us to find the patterns that work to hire candidates that self-identify as men/women respectively is particularly strong. But our Data Exchange customers - those who directly provide the hiring outcomes into Textio's engine to improve the system - get insights on all kinds of demographic attributes. Today the set includes gender, ethnicity, age, veteran status, and disability status. All of these will make their way into our core experience as the data sets are strong enough. If we're going to let Textio make recommendations on something as important as bias, we need to make sure they are statistically robust.
Andrew Ettinger@andrewett · PMM @Twitter // Previously @ProductHunt
How can NLP be useful to a founder?
Kieran Snyder@kieransnyder · CEO, Textio
@andrewett I have always been obsessed with quantitative approaches to language. For me, the form this took in school was a focus on NLP. I studied linguistics and math before I knew that would become a powerful combination in the tech industry. Right now this is a pretty powerful area of focus for tons of entrepreneurs. Learning loops are taking over the enterprise. Bots are hot. Virtual assistant space is exploding. But beyond the specifics of your business, people who study NLP tend to be good at math and good at language. That's a powerful combo for anyone in tech. We build tech, but work with people in order to do it.
John Fowler@johnsteerfowler · Founder, Zero Host
Hi Kieran. Thanks for doing a live chat. How do you think the recruitment/job postings industry is keeping up with technology? It's great to see companies like yours, but in general I feel it's still quite an archaic industry outside of Silicon Valley.
Kieran Snyder@kieransnyder · CEO, Textio
@steerj92 I come from a tech background and have been building engineering and product teams for many years. It's been eye-opening to me to see how hiring works in other industries, as it's often quite different. Our first customers were in tech and finance, which is not surprising when you consider how attuned these industries are to using data to make decisions in general. As we've expanded into retail, the arts, pharma, healthcare, manufacturing, etc - very different cultures for hiring. One thing that we get asked on customer calls all the time is what machine learning is. It's important to realize which context is shared and which context isn't...
Cynthia Awwsome@cynthiaaw
How can a computer program predict what someone will think when they read a document?
Kieran Snyder@kieransnyder · CEO, Textio
@cynthiaaw A computer program can't predict what someone will think, but it can predict how someone will respond. At Textio, we base these predictions on the responses to millions of other docs. As you're writing, your language is compared to the language of other similar docs with known outcomes, and the platform predicts how well your doc will perform based on similarities and differences.
Hi Kieran - You've written a lot about different gender styles in the workplace, as well as being a female CEO. As we approach a new phase of tech's relationship with the political system - and especially the new White House administration - what role do you think female CEOs need to play?
Kieran Snyder@kieransnyder · CEO, Textio
@sams18 All CEOs need to play a role, regardless of their demographic background. We all have the same job: building amazing, successful companies with great teams. Speaking only for myself, I take my responsibility to mentor and help other women in the industry (and other girls who want to go into the industry) very seriously. We need strong companies that are led by strong women - every one that exists makes it easier for the next person. We need to normalize strong female leadership - and help the next set of founders get there a little more easily.
Kieran Snyder@kieransnyder · CEO, Textio
@sams18 I think tech is still figuring out its relationship with the incoming administration. You can see it in the summit that just happened + in the reactions to the summit that just happened. Many tech leaders I know were disappointed with the election outcome, but have not yet worked through what course of action they should take. I see a lot of ambivalence and confusion. As a woman leading a company, mostly it makes me determined to make sure my company succeeds - and is a place where a variety of people enjoy working. You speak your political position, but in the end are accountable for the part that is clearly in your power to lead.
Eric Koester@erickoester · Insatiably curious.
Job postings are one piece of the hiring process -- do you have any insights about email message text or other sorts of "active" outreach activities? Do you analyze the other types of language used in the entire recruiting process (if companies are consistent?)
Kieran Snyder@kieransnyder · CEO, Textio
@erickoester I love this Q! Textio actually has a beta specifically for recruiting email. The language patterns that work in recruiting mail have some overlap with those in job posts, but there are many, many differences. The important thing for us in building our platform is that it can model any kind of writing where you're trying to get an outcome. The specific guidance that you'd give for writing a good job post is different from the guidance for e.g. writing a good email to your boss or writing a good marketing page. But the platform that can produce the right insights is shared across all. For your question about recruiting email in particular, some quick hits: Don't forget the subject line - both length and content matter. If your mail isn't open, it can't be read. The more you sound like every other recruiter, the less likely you are to get a response. But go try the beta and see what you think!
Kieran Snyder@kieransnyder · CEO, Textio
@erickoester I realized you also asked about whether companies are consistent. I firmly believe that software is one of the only things that can keep this consistency. People are not very good at following checklists and guidelines, even with the best of intentions. And not everything can be captured with guidelines!
Eric Koester@erickoester · Insatiably curious.
@kieransnyder that is 100% the magic of software... we can force constraints, test, learn and iterate. Very cool stuff... I'll dig in further and will reach out soon. Rock on.
Vinayak Ranade@pseudovirtual · CEO at Drafted
Hi Kieran - First, I'm a fan of Textio, great application for NLP+ML. I'm wondering if you are going to expose any of your excellent tech through free or paid APIs or partnerships for other startups (like us!) to leverage / hack on?
Kieran Snyder@kieransnyder · CEO, Textio
@pseudovirtual Thanks Vinayak! We don't have plans to expose our core predictive models via API at this time, though we may pursue exposing the scores themselves to other workflow components in the talent pipeline. Good luck with your startup! It is a journey.
Emily Hodgins@ems_hodge · Community and Marketing, Product Hunt
What’s the most common mistake textio pinpoints in job postings that aren’t attracting the right candidates for certain roles?
Kieran Snyder@kieransnyder · CEO, Textio
@ems_hodge Textio exposes millions of combinations that have statistical impact on job post performance, so it's hard to reduce it down to a simple list of tips. In general though, job posts suffer most when they aren't written for the purpose at hand. Typically what happens when someone opens a new position is they go find some other old job post and they try to modify it - they pull some bullets from this one listing, and the intro from another one... the resulting posts are not coherent. Many of the specific problems in people's listings come from this habit.
Cynthia Awwsome@cynthiaaw
Do you have jobs open at Textio now?
Kieran Snyder@kieransnyder · CEO, Textio
@cynthiaaw Yes! We are hiring, especially engineers. You can check out our open roles at https://textio.com/careers. Our job listings are outstanding! We just opened a VP of Sales role which is a really critical executive position for us, and I'd love to talk to people who are interested.
Ben Martens@martensben
Why would a company use Textio instead of just using a consulting firm to help improve their job listings?
Kieran Snyder@kieransnyder · CEO, Textio
@martensben Textio exposes millions of patterns, and the patterns that work in job posts change all the time. There are so many more patterns than even the best writer or consultant can keep straight in their head. One of my favorite examples of language changing over time is the phrase "big data." A few years ago, including "big data" in an engineering job post made jobs enormously popular. The term caught on, became popular, and it lost its differentiating effect - it's not differentiating if everyone has it. Then the hype cycle for "big data" went even further: it became so popular that it is a cliche. Now to include "big data" in your engineering job post actually exerts a negative effect on its performance. One thing I like to do when companies ask this question is point them to the Textio Index. Take the companies that have used this approach and look them up at https://textio.com on the homepage to see their scores. How well do they score?
Steve Richmond@sriche · CEO Potential+
Hi Kieran, I really think what Textio is doing is interesting and congrats on all the great progress. Job postings are generally posted publicly. Recruiting emails go out broadly. Will textio continue to focus on communications that are more public in nature, or do you think you'll start to turn your attention to more personal communications, such as person-to-person email, etc?
Kieran Snyder@kieransnyder · CEO, Textio
@sriche Thanks Steve! Textio's platform attaches anywhere 1. You have enough document data to be statistically interesting and 2. You have measurable outcomes for all the documents in the data set, so you can find the patterns that "work." The difference between public comms and more private comms is fundamentally just about the outcome you're looking for. With person-to-person email, you have one goal: get the other person to read + respond as appropriate. With public comms, there many be many goals, but in most cases you are looking to get lots of people to take some kind of action (apply for a job, sign up for your product, whatever). Textio is interesting anywhere you can capture those outcomes. =)
Kieran Snyder@kieransnyder · CEO, Textio
Thanks so much for all the great questions today, everyone! Check us out at http://textio.com/ and look yourselves up in the Textio Index to see how your jobs perform. Talk soon!