When RIFs Have Different Impacts
Two big companies announced job cuts within three days of each other this month. American Express went first, on October 23. Microsoft followed on October 26. Both came with the same backdrop: large companies trimming staff while they pour money into AI.
From the outside, they look like the same move. I'm covering them together because they put weight on different parts of each company. If you lead a team, that matters. It means the repair has to be different too.
The two moves
Amex is cutting about 3,500 jobs. That's roughly 5% of its 78,000 people, and its first big round of cuts in about a decade. It came in a quarter when net income rose 28%. The CEO said he'd asked senior leaders earlier this year to streamline their groups and cut layers so decisions move faster. Amex hasn't said which units or levels are affected. But the company chose. Leaders drew the lines and picked the roles.
Microsoft tried something new. It offered about 60,000 people, roughly 27% of its workforce, the chance to leave with a package close to its normal severance. The offer covers people at vice president and below in the US, Canada, India, and the UK. Some groups are left out, including parts of data centers, sales, and marketing. People have until early December to decide. Anyone who says yes leaves on December 31. Its people chief told staff, "The Voluntary Exit Offer is new for us. We'll learn from it."
So here's the split. At Amex, leaders pick who goes. At Microsoft, employees pick themselves.
A quick tour of the Bridge
Linden's Bridge is the model I use to see where pressure lands in an organization. It has five layers. From the bottom up, they are Awareness, Purpose, Rules, Systems, and Execution.
Awareness is how an organization notices what's coming. Purpose is what it optimizes for when it has to choose. Rules is where it settles who gets to decide. Systems are the tools and handoffs that let the work happen. Execution is where the load becomes visible.
Load climbs. When a layer stops carrying its share, the weight doesn't vanish. It moves up until something visible gives. That's why the place something breaks is rarely the place it started.
Amex: a planned change to Rules that loads Systems
Amex was clear about what it wanted to change. It hasn't said exactly where the cuts fall, but the CEO pointed at layers. Fewer layers means fewer people between a question and an answer. That's a deliberate change to Rules. Decision rights move closer to the work.
That part can work. Here's the catch. Middle layers rarely just decide things. They also carry handoffs. A mid-level manager is often the person who turns a strategy deck into this week's priorities. She knows which team to call when a customer issue crosses two departments. She catches the thing nobody owns.
None of that shows up on an org chart. So when the layer goes, the decision rights get redrawn on purpose. The handoffs don't. They land on whoever's nearby: the manager above, who now has twice as many direct reports, or the frontline lead, who now has to guess.
Picture a customer dispute that used to take two days because one manager knew who to loop in. Now it bounces between teams for a week. Nobody made a bad call. The handoff just left with the role.
So at Amex, Rules changed by design. Systems takes the strain by accident.
Microsoft: a hands-off rule that loads Awareness
Microsoft's move puts weight somewhere else. It handed the "who goes" decision to the people themselves. That's also a Rules change. But the bigger effect is on Awareness.
For the next five weeks, Microsoft can't know what its own structure will look like on January 1. The company said similar programs elsewhere saw about 5% to 6% of people take the offer. Its own early retirement offer this spring drew about 30%. Apply those rates to 60,000 people and you get anywhere from about 3,000 departures to about 18,000. The company won't know which ones until December.
Self-selection also has a pattern. The people most likely to say yes are the ones with good options elsewhere. Those are often the people their peers lean on most. One team of ten might lose nobody. The team next to it might lose the two people who understood the oldest code in the building.
Leaders can't plan for that the usual way. They won't know where the gaps are until people decide.
The offer does carry one strong Purpose signal. The groups Microsoft left out tell you what it won't risk. Nobody in those parts of data centers, sales, and marketing gets a package to walk out the door. That says more than the memo does, because it shows what the company protects when it has to choose.
What breaking would look like
My model says both will show up first in Execution. That's the top layer, and it covers for the layers below until it can't.
At Amex, a break would look like slower service. Customer issues that cross departments would take longer. Managers with wider spans would miss things. The cause would sit a layer lower, in Systems, where the handoffs went missing.
At Microsoft, a break would look like a team that stalls early next year. A product slips. A fix takes three weeks instead of three days. The cause would sit lower still, in Awareness. Nobody could see in advance which knowledge was walking out.
Same kind of move. Different crack.
Why the rebuild is different
This is the part that matters if you're running a team.
If you cut the way Amex did, you chose who goes. That means the repair can happen before anyone leaves. The fix lives in Systems, and you have time to build it.
If you cut the way Microsoft did, you can't start there. You don't know yet who's leaving. So the repair has to start in Awareness: first find the holes, then fill them. The fix still ends in Systems.
Swap the repairs and both miss. A handoff map built in October won't help Microsoft, because the people on it may be gone by January. Waiting until December to see who left won't help Amex much, because Amex already knew who was leaving. It just lost the chance to plan.
Lowering the risk
The leader who made the call owns this part. It can't wait for someone below to raise it. An employee who asks "did anyone write down what Maria did?" is really asking whether the boss planned the cut well. Most people won't ask that out loud. So the check has to be built into the cut itself.
If you picked who goes. Before each person's last day, their manager sits down with them for an hour. Together they list what that person did that isn't in the job description. Who called them for help? What did they catch before it became a problem? Which handoffs ran through their desk? Every item gets a new owner by name, and that owner hears about it before the person leaves. Then the leader watches one number for the next 90 days: how long it takes to close an issue that crosses two teams. If that number climbs, a handoff didn't land. You'll see it there months before you see it in customer complaints.
If people picked themselves. You can't protect knowledge you can't see yet. So shrink the blind spot now. Before the window closes, ask each team lead privately to name the people their team would struggle most to lose. That list isn't for talking anyone out of leaving. It tells you where to look first. When the yes answers come in, compare them to the list. Microsoft's people decide in early December and leave December 31. That leaves a few weeks where the person is still in the building. Wherever a name shows up on both lists, start the handoff work that week.
One check works for both. Look at what the cut left alone, and compare it to what you told people matters most. If the two don't match, your staff has already noticed. Say plainly what you protected and why, before the gap does the talking for you.
If you're leaving, or deciding whether to. Look at what the cut protected. That's the clearest Purpose signal a company gives, because it shows what leaders actually care about when it matters. At Microsoft, parts of data centers, sales, and marketing weren't offered the package at all. If your work sits close to what got protected, the company is telling you where it plans to keep spending. If it sits far away, that's worth weighing too.
If you want the company to succeed, write down the stuff you do and the handoffs you make that aren’t in the job description. Give it to whoever takes over, even if nobody asks. It costs you an afternoon. It's also how the people you leave behind will remember your exit, and some of them are likely to be your references.
If you're staying. All the weight lands on you. Execution is where the Bridge shows strain. When something starts taking longer, check whether a handoff left with someone before you blame your own workload. Then raise it as a structural gap, a missing handoff. "Customer escalations used to go through Dana. How are they handled now?" That's a question about Systems, and it comes with its own fix. A manager can answer it without having to admit a mistake, which makes it more likely to get answered.
A cut is meant to save money. Saving money requires that operations continue. Every ball dropped costs the company time and opportunity. The structure that gets built to accommodate the staff reduction is what determines how much of the potential savings the organization actually keeps.
