Layoffs · Primly Community

post-layoff job search data: my 11-week pipeline with actual conversion rates

ds_dmitri · 4 replies

data scientist brain here, couldn't not track everything. laid off in april 2026, accepted an offer 11 weeks later. here's the full funnel.

applications by source: linkedin easy apply: 44 applications, 3 phone screens (6.8%) company career sites (direct): 28 applications, 8 phone screens (28.6%) referrals: 9 applications, 7 phone screens (77.8%) recruiter inbound: 6 outreaches, 5 phone screens (83.3%)

not a surprise that referrals and inbound dominate. the surprise was how badly easy apply performed. i used it for convenience early on and it was nearly worthless.

stage-to-stage conversion: application to phone screen: 26.7% (skewed heavily by source) phone screen to technical: 68% technical to onsite: 52% onsite to offer: 33%

the weakest point was onsite to offer. i failed two onsites before i started doing more systematic post-mortems. the pattern: my SQL and stats rounds were fine, my case study rounds were weak. i spent weeks 7-8 specifically reworking how i present model selection decisions to non-technical stakeholders. that changed things.

time spent per week: applications: 8-10 hours prep: 12-15 hours (heaviest in weeks 5-8) networking/coffee chats: 3-5 hours

total offers: 2, one DS senior role at a fintech, one at a mid-size healthcare SaaS. comp difference was meaningful (~$22k TC difference). took the lower TC offer for the team and scope, don't regret it so far.

one honest thing: week 6 was brutal. every active process had stalled or ghosted. if you're in week 6 and wondering if something is wrong with you, it might just be the timing. two of my best processes restarted in week 7 with no explanation.

4 replies

ops_omar

the referral conversion number (77.8%) is almost too good to be normalized for selection effects right? like referrals are self-selecting for people who know someone relevant to the role. still, the directional point is true. my ratio on direct applications was around 22% and referrals was closer to 60%.

ds_dmitri

100% selection bias. i wasn't requesting referrals randomly, i targeted roles where i had a genuine connection. so the 78% number is really 'quality of referral' as much as 'value of referral mechanism.' the actionable version is: don't burn a weak tie referral on a role where the connection barely knows you.

staff_steph

the case study round insight is real. DS/MLE candidates who can't translate model decisions to business terms fail at the debrief stage all the time. it's rarely the technical content that sinks you at that level, it's the framing.

market_realist

week 6 brutal resonates so much. i'm in week 9 and just got two rejections this week after silence. good to know it's not always a straight line down.