Kastle's 66% Tuesday Is a Sourcing Calendar. Send Wednesday at 7am.
Kastle's 66% Tuesday badge peak plus Ma and Ding's RTO attrition research point to a 6 to 8 week sourcing window for senior engineers. Here is how to work it.
Analysis Atlas's May 14 deep dive stapled three numbers together that most recruiters are still reading separately: Kastle's 66% Tuesday badge peak, AAA's $11,577 annual cost of owning a car, and the Ma and Ding finding that time to fill stretches 23% at firms after a return to office mandate. Read together, they describe a specific, time boxed sourcing opportunity. The cohort is long commute senior ICs. The window is six to eight weeks. The opener writes itself if you let the data write it.
This is not an RTO opinion piece. It is a sourcing playbook.
The Tuesday peak is a calendar, not a real estate stat
Kastle Systems' Back to Work Barometer pulls keycard and fob data from 2,600 buildings and 41,000 businesses across major U.S. metro areas. In the December 2025 reading, the 10 city average hit 56.3% for the week, a new post pandemic high. The Tuesday peak crossed 66.0% across all building types and 95.5% in A+ rated buildings. Friday, by contrast, sits roughly 43 points lower.
Real estate analysts read those numbers and talk about lease renewals. Recruiters should read them as a weekly emotional curve. Tuesday is when the parking garage is full and the train is standing room only. Wednesday morning is when a senior engineer who left her house at 5:45am Tuesday opens her laptop, sees a calendar packed with the same meetings she could have taken from her kitchen, and is most receptive to a message that is not generic.
That is the calendar. Monday night and Wednesday morning are when commute pain is psychologically rawest. Friday afternoon outbound, the default of most agency tooling, is the worst possible time to land in this cohort's inbox. They are already mentally home.
Why Ma and Ding matter more than any quit survey
Mark Ma and Yuye Ding at the University of Pittsburgh's Katz Graduate School of Business looked at 54 S&P 500 firms that implemented RTO mandates between April 2020 and June 2023, against three million LinkedIn profiles. The headline finding is a 14% jump in departures versus pre RTO baselines. The useful finding, for sourcing, is downstream of that.
Time to fill at the mandating firm increased 23%. Hire rate dropped 17%. That is your window. While the RTO firm scrambles to backfill, its remaining senior ICs are exposed for roughly six to eight weeks before either a counteroffer culture sets in or the strongest performers have already quietly accepted somewhere else.
The cohort skew is sharper still. Ma's quote: "The probability of more skilled employees departing after RTO mandates is 77% higher than that of less skilled workers, and the probability of senior employees departing after RTO mandates is 36% higher than that of junior workers." Women, seniors, and high skill workers leave at disproportionately higher rates.
These are not the people posting "open to work" frames on LinkedIn. They update a headline. They accept a conference talk. They push to a side project repo they had not touched in a year. The behavioral signals are there, but Boolean queries do not catch them.
Defining the super commuter, precisely
UCLA's Institute of Transportation Studies uses the formal definition: a one way commute over 50 miles or 90 minutes. Their work on greater Los Angeles found super commuters are a small but growing share of workers who skew higher income than other commuters. That is the senior IC profile you want, by a different name.
Cross reference this with the AAA 2025 figure: $11,577 per year to own and operate a new vehicle, or $964.78 per month. That is before parking, before tolls, before the eight to twelve hours a week a 60 mile each way commute eats. The Census release also confirmed the count of workers with one way commutes over 60 minutes climbed again in the most recent reading.
So the filter is: senior or staff or principal engineer, home address more than 50 miles or 90 minutes from a known RTO mandating HQ, employed at a firm with a five day mandate effective in the last twelve months. That is a tight query, and it is not one LinkedIn Recruiter's UI was built to express. Describing the person in plain English and getting back a ranked shortlist across GitHub, LinkedIn, and the open web is the entire reason we built Refolk: "senior backend engineers in the Seattle metro who live more than 50 miles from downtown and work at Amazon" is a sentence, not a Boolean.
The geographic overlap nobody is mapping
Refolk's index shows roughly 239,000 U.S. based Senior, Staff, and Principal software engineers, concentrated in NYC, the SF Bay, Seattle, Chicago, Austin, and Boulder. Those are the same metros where Kastle records the highest Tuesday peaks. The sourcing pool and the RTO friction overlap geographically, which is what makes this a workable strategy and not a thought experiment.
Amazon, Dell, Disney, and Walmart are the most cited five day mandate firms in 2025 and 2026 coverage. Cheryl Ainoa, Sam's Club's former CTO, resigned rather than relocate to Bentonville after Walmart's RTO push. That is one named, public, senior technology departure in the exact pattern Ma and Ding describe. There are hundreds of quieter ones happening right now in the layer below CTO, and those are the people you can actually hire.
The destination side of the message matters too. Nvidia and Airbnb continue to operate in flexible or remote modes. Naming a specific employer brand that has not flinched on flexibility, instead of pitching "work life balance" in the abstract, is what gets the reply.
The first message, written from the data
The Barrero, Bloom, and Davis work pegs the value workers place on a two to three day work from home arrangement at 5 to 7% of pay, with higher figures for women, parents, and high earners. The AAA number monetizes the commute side. Combine them and your opener stops sounding like every other recruiter in the inbox.
A subject line like "$11,577 plus 8 hours a week back" outperforms "exciting opportunity at a Series C." The body should reference the candidate's actual commute distance, the specific RTO policy at her current employer with date, and a destination role that explicitly states the in office expectation in days per week. Three sentences. No "I came across your profile."
The flight prone cohort does not post open to work. They update a headline, accept a talk, and push to a repo they had not touched in a year.
A 2025 CNBC survey found only 40% of employees said they would quit or look for another remote job if given a mandatory RTO notice, down from earlier readings. The tighter labor market means volume strategies fail. The 40% who would move are higher quality than the historical base rate, but you have to find them individually. This is the part where most agencies still send 400 InMails and call it a campaign.
Timing the send
If Kastle's badge curve is your calendar, the send schedule is:
- Monday 8pm to 10pm local, when the next morning's commute is looming and the candidate is on her phone.
- Wednesday 6am to 8am local, after the Tuesday peak, when she is either back in the car or staring at a kitchen counter trying to talk herself into it.
- Avoid Friday afternoon entirely. The cohort is mentally checked out of work and your message gets archived with the weekly newsletters.
This is not folk wisdom. It is the badge data inverted. The Tuesday peak tells you when the pain is acute. The send goes immediately before and immediately after.
What to put in the screen
Once you have a reply, the screen has to honor the premise. If you opened with commute math, the first call needs to confirm the destination role's in office expectation, the home metro, and the commute the candidate would actually have. Vague language about "hybrid" kills the trust the opener built. Three of the named flexible employers, including Nvidia and Airbnb, publish specific weekly in office expectations. Match that specificity.
A useful pattern on the screening call: ask the candidate to dollarize her current commute herself. Most have not done the math. When she says the number out loud, the offer stage gets easier.
A note on Boolean and why it is breaking here
The Ma and Ding cohort, women, seniors, skilled ICs, does not raise a hand publicly. They do not change their LinkedIn to "open to work." They edit a skill, accept a CFP, push a commit, follow a recruiter back. The signals are real but they are scattered across GitHub, LinkedIn, conference sites, and personal blogs. Stitching them together with x ray search and a stack of Chrome extensions is what sourcers have been doing for a decade, and it does not scale to the six to eight week window the RTO data opens.
This is the second place Refolk earns its keep on this play. Passive candidate sourcing in 2026 is less about clever Boolean and more about expressing intent in plain English and letting the system reconcile signals across surfaces. "Senior infrastructure engineers at Amazon Seattle whose GitHub activity picked up in the last 90 days and who have not changed jobs in over four years" is a query that took an afternoon of manual stitching last year. It should take a sentence.
The window closes
Time to fill stretches 23% at the mandating firm. That is roughly the gap between when the senior IC starts thinking about leaving and when her manager realizes the bench is thin. After that, counteroffers harden, retention bonuses get cut, and the strongest performers either leave or get bought back. Recruiters who treat the Kastle reading as a real estate story will keep sending Friday afternoon InMails. Recruiters who treat it as a calendar will be twenty replies deep by the time the rest of the market notices the cohort is in play.
The 66% Tuesday peak is not a story about offices. It is a story about which Wednesday morning you send.
FAQ
How do I identify super commuters without home address data?
You do not need exact home addresses. The UCLA definition is 50 miles or 90 minutes one way, and home metro plus employer HQ is enough to estimate that at scale. LinkedIn location fields, GitHub profile locations, and conference bios usually surface the metro. Cross reference with the employer HQ and a 50 mile radius and you have a workable filter. Tools that combine these signals across sources, including Refolk, do the reconciliation automatically.
Does the Ma and Ding research apply outside the S&P 500?
The study covered 54 S&P 500 firms because LinkedIn data was reliable at that scale, but the mechanism, senior and skilled workers leaving disproportionately after RTO mandates, is not specific to large caps. Reporting through 2025 and 2026 has found the same pattern at mid market firms and at private companies that mirrored Amazon's five day policy. Treat the 14% departure jump and 23% time to fill stretch as directional, not as ceilings.
Is the $11,577 number safe to put in a first message?
Yes. It is the AAA 2025 figure for total annual cost of owning and operating a new vehicle, $964.78 per month, down $719 from 2024. Cite AAA explicitly. Candidates check, and the source holds up. The number resonates because most people have never added insurance, depreciation, and maintenance together. Saying it out loud does the work.
What if the candidate's current employer counters?
Counters are the main reason the window is six to eight weeks and not six months. Two defenses: lead with the role's in office expectation, not just compensation, because money counters are easier to write than policy counters, and screen for candidates who have already had the commute conversation at home. If a partner or a school run is in the picture, the counter rarely sticks. If it is pure compensation, expect to lose roughly a third at offer stage and price your pipeline accordingly.