eighteen month writeup on transitioning our anxious shepherd mix from four years of pure work from home to a hybrid three days in office schedule, the eight month protocol that actually worked, and the two things the popular separation anxiety blogs got wrong for our dog

Wren is our 6 year old female shepherd mix, 48lb, we adopted her at 18 months in the middle of 2020 right when our company went permanent remote. she had a rough early history (returned to the shelter twice before us, second return cited "destructive when left alone") and the four pure-WFH years were genuinely good for her, she essentially followed me from room to room from 2020 through 2024 and rarely spent more than 2-3 hours alone. in summer 2024 my company announced a hybrid mandate of 3 days a week in office starting january 2025, and we knew this was going to be a real project given her history. we are now 18 months past the start of that work and she is in a stable place doing 6-7 hours alone consistently with no apartment damage, so i want to share what we actually did because the version of the protocol that worked for us is meaningfully different from what the popular separation anxiety blogs recommend.

the starting baseline at month 0. she could do about 90 minutes alone before she started panting, pacing, and chewing the door frame near the front door. we had hard evidence from a camera we'd set up about 18 months earlier when we tried a single 4 hour absence and came back to a chewed door frame, a urine puddle (she is fully house trained), and a clearly elevated dog. so we knew exactly the size of the gap between where she was and where she needed to be (4-6 hours of unsupervised alone time at low stress).

the eight month protocol we built and stuck to. one, months 1-2 we did relative-calmness work at home, not absence training yet. the goal was to build her ability to be in a different room from us with the door open, then with the door closed, for increasing durations. lots of frozen kong work in the second bedroom while we were on the couch. by end of month 2 she could be in a different room for 30 minutes confidently. two, months 3-4 we started actual absence training with sub-30-second absences (open the door, step into the hallway, step back in, treat). built duration over 8 weeks to 8-10 minute absences by end of month 4. camera always on, sessions paused if she showed early signs of stress. three, month 5 we standardized the morning routine that would eventually be the real work routine, same wake time, same walk pattern, same enrichment toy stuffed the night before. the predictability piece matters because dogs predict absence from contextual cues and standardizing the cues means the eventual absences feel less novel. four, months 6-7 was the unlock for us and where we deviated from the published protocols, my partner who works from home permanently did "fake leaves" where he would go through the full departure routine (keys, jacket, door close, leave), then sit silently in his car in the garage for 30-90 minutes while i monitored the camera, then come back through the front door at a normal time. this built the bridge between "you are home but i can't see you" and "you are actually gone." we did this 3-5 days a week for 8 weeks. five, month 8 i started doing my actual 3 office days per week with neither of us home, by which point her alone-tolerance was 4-5 hours with no damage and no escalation on the camera. we added doggy daycare 1 day a week (one of my 3 office days) to break up the week. six, months 9-18 has been maintenance and gradual extension. she's now confidently doing 6-7 hours alone with no damage. the camera shows mostly sleeping with 1-2 short bouts of restlessness around lunchtime.

what we did NOT do that the popular SA blogs recommend, and why. one, we did not do the malena demartini sub-second graduated absence protocol that goes for 6-12 months at the longer end. that protocol is genuinely the right answer for true clinical separation anxiety where the dog cannot tolerate the human moving out of sight for 5 seconds without panicking, and if Wren had presented that way we would have used it. she did not, she presented with learned dependency on our constant proximity over 4 WFH years with a baseline alone-tolerance of 90 minutes that we needed to extend to 6-7 hours. for that profile we could move faster than the SA protocol. the framework we used (which our trainer suggested) was distinguishing "context-dependent alone-distress" from "true SA," and the former responds to context-rebuild + gradual duration extension, the latter requires the sub-second protocol. our trainer was a CSAT and was very clear about which one Wren had. two, we did not put her in daycare 5 days a week, which was actually our first instinct. our trainer pushed back hard on this, daycare 5 days a week builds a different dependency (on daycare) and does not actually develop the home-alone skill that we needed. once a week daycare as part of the maintenance pattern, yes. as a substitute for the alone training, no. three, we did not start meds first. we discussed fluoxetine with our vet at month 1 and made the decision to save it as a fallback if the behavioral work plateaued. it didn't plateau so we never used it. i want to be clear that for other dogs and other families meds may be the right first step, this is not an anti-meds writeup, it is a "we built our specific protocol with our specific dog and meds were not necessary in our case" writeup.

the thing that was nontrivial that i would do again. the partner-fake-leaves phase in months 6-7 was the unlock. most published protocols don't include this step explicitly, they go from "you are home but in another room" directly to "you are gone." for Wren that gap was too big. the two months of fake-leaves where one human was technically home but completely absent from her experience (silent in the garage, no scent path from her crate) gave her the bridge between "out of sight in the house" and "actually gone from the building" without the discontinuity that had caused the original failure. our trainer said about 30% of her clients need this bridge phase and the rest can skip it. happy to answer specifics on any of the months if useful