Separating work and personal browsing: the setup order that holds up

Most attempts to separate work and personal browsing on a Mac fail in the same way. The profiles get created on a Saturday, the colours get picked, everything looks clean, and by the following Friday the personal Gmail is back in the work window because that window happened to be in front. Nothing was misconfigured. The setup was simply built in an order that could not survive a busy day.

The order matters more than the settings, because settings can be changed later and decisions cannot. What follows is the sequence: three things to decide before opening any preferences pane, then five steps that each depend on the one before it.

Decide three things before touching a settings screen

These take about thirty minutes with a pen. Skipping them is what produces the setup that gets abandoned.

Count identities, not services

The common mistake is counting services. Mail should be separated, then chat, then documents, then the client's ticket tracker, and the list grows until nothing is manageable.

The number that matters is how many identities exist. Someone employed full time who also takes freelance work has three: the employer, the freelance business, and private life. Someone working across three client organisations with separate credentials for each has at least three, possibly four. That count becomes the number of storage areas needed. Services do not enter into it, because one identity contains all of them: its mail, its chat, its documents, its files.

Decide which side owns the default browser

macOS allows exactly one default browser. Every link clicked from Mail, from Calendar, from a native chat client, from a PDF, opens there regardless of which identity that link belongs to. Deciding this after the split is built means every wrong-side link costs a reopen, and that cost is what erodes the habit.

The test is simple: which side generates more clicks that start outside the browser. For most employed people that is the work side, because meeting invitations, shared documents, and ticket notifications all arrive in native apps. The other side then gets entered directly, through a bookmark or a fixed window, not through links.

Pick one trigger for switching

A split survives when the trigger is single and automatic. Time is one option: everything after 18:00 is personal. Place is another: the desk is work, the sofa is personal. The tool itself is a third: opening a particular window means being on that side.

Two triggers is the same as none. Every switch then requires a small decision, and the side that requires a decision is the side that stops being used.

Step one: create exactly as many storage areas as identities

Now the settings screen opens. In Chrome and Edge this is a profile, in Safari a profile, in Firefox either a profile or a container from the Multi-Account Containers extension. A profile keeps its own cookies, history, passwords, and extensions, which is precisely the boundary that stops one identity's session from being used by another. Google documents profiles as being for exactly this case, sharing a computer or keeping work and personal accounts apart.

Create the number decided in the first section. No spares. An unused storage area does nothing except lengthen the switcher menu, and a long switcher is one more decision per switch.

One thing has to happen immediately after creation: sign in to each storage area with only the accounts belonging to that identity. Adding one personal account to the work profile because it was convenient at the time removes the boundary entirely, and the failure will not be visible until something is sent from the wrong address. Password managers need the same treatment. If the manager supports separate vaults, bind one vault per storage area rather than unlocking everything everywhere.

Two edge cases come up at this point and both are worth settling now rather than later. The first is the account that genuinely belongs to more than one identity, usually a shared client tool or a subscription used for both work and private reading. Put it on the side where it is opened most often and reach it from the other side through a plain bookmark. Duplicating the login across storage areas defeats the purpose, because the moment two areas hold the same session, the boundary stops being verifiable by looking.

The second is two-factor authentication. Codes come from an application on the device, not from inside the browser, so that layer stays shared no matter how many storage areas exist. Naming each entry in the authenticator after the identity rather than the service removes the pause that otherwise happens every time a code is needed.

Containers behave differently from profiles and the difference is worth knowing before choosing. A container belongs to a tab, so two identities can sit in one window as two tabs. Every profile-based approach needs a second window for the second identity. Neither is better in general. Containers keep the window count low, profiles keep extensions separate as well as cookies.

Step two: put the marker at the edge of the window

A storage area with no visible marker is a boundary that has to be checked by reading. Chrome profiles take a theme colour and an icon, and both should be treated as instruments rather than decoration.

The marker belongs at the edge of the window, not in the content. Content changes with whatever page is loaded. The edge does not. A colour visible in peripheral vision removes the act of checking, which is the actual cost being eliminated here, not the act of switching.

Splitting web apps out as separate applications does the same job more strongly. Each one lands in the Dock with its own icon and appears separately in the application switcher, which removes tab hunting from the process altogether. The Features page covers how a single window can hold many web apps while still being addressed individually.

Step three: give every window a fixed address

Markers tell which side a window belongs to. Addresses tell where it lives. macOS handles this through Mission Control, which creates additional desktops called spaces.

If the desktop on your Mac gets cluttered with open app windows, you can use Mission Control to create additional desktops, called spaces, to organize the windows. When you work in a space, only the windows that are in that space are shown. Source: support.apple.com

Apple's limit is 16 spaces. Two is enough for a work and personal split, and more than four makes the horizontal swipe long enough that finding a window becomes its own task again.

Assignment happens from the Dock. Control-click an application icon, choose Options, then choose under Assign To:

Assign To option	Where the app opens
All Desktops	In every space
This Desktop	Only in the current space
Desktop on Display [number]	In the current space on a specific display
None	In whichever space is active at the time

Work browser to the work space, personal browser to the personal space. One default behaviour catches people out here: switching to an application automatically moves the desktop to a space that already has windows open for it. That is why a new window sometimes appears to teleport to the other screen. The setting lives in System Settings under Desktop & Dock, in the Mission Control section, phrased as switching to a space with open windows for the application.

Give each space a different wallpaper while setting this up. The point is identical to the window marker: make the current side visible without reading anything.

Step four: cut the number of entrances to one

At this stage the boundary exists. What breaks it is having several ways in. If the work side can be reached from a Dock icon, from Spotlight, and from whichever window happens to be frontmost, then each of those lands somewhere slightly different, and under time pressure the nearest one wins.

Pick one entrance and remove the others from reach. A fixed Dock position or a single keyboard action both work. What does not work is keeping two and intending to use one. Removing the unused icon from the Dock takes ten seconds and prevents the most common form of collapse. Keyboard-driven entry to a specific window, of the kind described on the Built-in terminal page, follows the same principle: one action, one destination, no aiming.

Step five: draw the notification line

Notifications are last because they are the only layer that keeps working when the others are half built, and the only one that cannot be fixed by the others.

Focus modes on macOS select which applications may deliver notifications, and they can switch on a schedule or by location. A work focus silences personal applications and a personal focus silences work ones. If the trigger chosen in the third decision was time, set the focus to change at that time, and the switch stops depending on memory.

Silencing everything rarely lasts. One missed message produces enough anxiety that the whole thing gets turned off within a week. Allowing three or four specific senders through and silencing the rest survives, because nothing important is being gambled.

What breaks after a week, and what it means

Collapse has recognisable shapes. Each one points at a specific step.

Symptom	Which step failed
The wrong window gets used when in a hurry	Step four, a second entrance is still reachable
Links keep opening on the wrong side	The default browser decision was never made
Notifications do not follow the switch	Step five is manual rather than scheduled
One side keeps getting signed out	An extension that handles sessions exists in only one storage area
The split feels pointless	Too many identities were created in step one

None of these require rebuilding. One step is broken, the rest are intact, and repairing one step takes a few minutes.

The other honest observation is that this whole sequence is free and also manual. Storage areas, markers, spaces, entrances, and focus modes each live in a different settings screen, and none of them know about the others. For two identities that is a thirty minute job that holds. For five or six, the number of separate settings starts to outweigh the benefit, which is the point at which a tool that treats those things as one unit becomes worth pricing. A comparison of how different tools weight session separation against window placement sits on the Compared with Sidekick page, and the cost side is on the Pricing page.

What to change first

Do not start with profiles. Start by writing down how many identities actually exist and which side owns the default browser, because every later step inherits those two answers. Build the storage areas next, then markers, addresses, entrances, and notifications in that order, and stop at the point where the remaining friction stops bothering you. If the manual version holds for a week and the only complaint left is the number of settings screens involved, that is the moment to look at SpaceDeck.