The Lottery Stopped Being a Lottery

For most of the electronic registration era, cap season planning was a volume question. Identify everyone plausibly eligible, register them, and accept that the outcome was a coin toss you could influence only by entering more coins. That planning model no longer describes the system.

For FY 2027 the process is no longer a purely random lottery. The final rule establishing the change was scheduled for publication at the end of December. What that does is move a set of decisions forward, out of the petition stage and into the months before registration opens.

What has not changed

Start with the fixed points, because they still govern the arithmetic. The annual numerical limit provides sixty-five thousand initial visas, with an additional twenty thousand reserved for workers holding advanced degrees from US institutions. Those numbers are set by statute and no rule about selection method alters them.

The registration mechanism is also intact. USCIS has used an electronic registration system to allocate these visas since March 2020, with employers initially supplying only basic biographic information for each candidate. Employers still register each beneficiary and still pay the registration fee of two hundred and fifteen dollars per registration.

So the shape of the season is familiar. A short registration window, a selection event, and a filing period for those selected. What changed is the part in the middle that everyone used to treat as weather.

Why a weighting changes the planning cycle

Under a flat random draw, the only lever available to an employer was the number of registrations. Nothing about how a role was structured altered its odds, so structural questions could be deferred until after selection, when they mattered for the petition itself.

Once outcomes vary by an attribute of the filing, those questions move to the front. The determination of how a position is classified and compensated stops being petition paperwork and becomes part of the selection strategy, decided while there is still time to change something.

That has organisational consequences well beyond the immigration function. The relevant conversations involve compensation, job architecture and hiring managers, and they need to happen on a schedule that ends before the registration window opens rather than beginning when it closes.

It also changes what a registration costs in real terms. Under a flat draw, an extra registration was a cheap option on a fixed probability, and the rational move for many employers was to register broadly. Where odds vary by filing attribute, breadth stops being uniformly efficient: registrations concentrated at one end of the distribution behave differently from the same count spread across it, and the fee is charged per registration either way.

What to do before the window

• Inventory the roles, not just the people. Build the candidate list against actual position descriptions and pay data rather than against a spreadsheet of names inherited from last year.
• Get compensation into the room early. Any decision touching wage level is a compensation decision with consequences beyond immigration, and it cannot be made unilaterally by counsel in March.
• Document the basis for every classification. Whatever position you take must be defensible at petition stage on the same facts. A classification chosen for selection reasons and unsupported by the role is a problem you have merely deferred.
• Model the season with ranges, not a single plan. Build the hiring plan around several selection outcomes, because the honest answer about how many of your registrations will clear is still a range.

Say what is knowable and stop there

The temptation in a changed system is to sell certainty. Resist it with candidates and with hiring managers alike. What can be stated is the published rule, the statutory caps, the fee and the mechanics. What cannot be stated is any individual's odds.

Give people the calendar instead: when registration opens, when it closes, when selection results appear, and what happens in each branch. A candidate who understands the sequence handles an unselected outcome far better than one who was told their profile made them likely.

And record the reasoning behind each registration decision at the time you make it. Rules in this area change between seasons, sometimes between quarters. A file that shows what the position was, how it was classified and why, on the facts as they stood, is the only durable answer to a question asked a year later by an auditor, a candidate or a hiring manager who remembers the conversation differently.