Four revenue streams feed one P&L. About two-thirds goes to running the business. What's left is EBITDA, and the board's question is how much of it survives capex, interest and tax to become free cash. In 2025 that was $2.2B.

Same three segments, two very different pictures. The network is most of the revenue and nearly all of the profit. Health and AI are more than a fifth of the revenue but, so far, a much smaller slice of the profit.

What it costs to win one subscriber, how long until they've paid it back, and what they're worth by the time they leave. It runs on the Q4 plan's channel costs: winback $180, in-market $420, flank $140, blended $276.

This is the number the board watches. Of every $7.35 of EBITDA, $2.20 became free cash in 2025. The 2026 plan takes that to $2.45B by spending less on capex, not by cutting the network.

Each company's 2026 EBITDA and interest bill, ticking at its reported pace. Each bar is its net debt on one shared scale, and the solid part is the EBITDA it has earned against it this year; the marker is where it lands on Dec 31. The gauges below show where each ratio sits today.

The core solvency metric. The teal marker on each bar is that company's stated deleveraging target.

Same destination (~3.0×), different arrival times. TELUS still gets there about two years ahead of Bell, but a year later than it planned.

A candidate who walks in with a billion-dollar shopping list fails the room. The board's priority is deleveraging. Credibility comes from acquiring surgically, funded by the monetisation levers already public.

Each strengthens a metric where you beat — or need to beat — the rivals. Sized to stay leverage-neutral.

Carriers don't publish a clean CAC line, so we model it from the drivers you do control. This is the math you'd defend in the boardroom — and why low churn beats cheap adds.

Particles show the net direction of port-ins and port-outs. Freedom (Quebecor) is the disruptor pulling price-shoppers from all three incumbents; TELUS's low churn means it leaks the least. Hover a carrier to isolate its flows.

Loyalty is the inverse of churn. Average customer tenure = 1 ÷ monthly churn rate. This is TELUS's single greatest structural advantage.

Two views side by side: the spectrum bands that power "5G+" (mid-band 3.5/3.8 GHz is the game-changer), and the actual median speeds customers measured.

Every mobile site licensed in Canada, from ISED's spectrum register (the same data celltowers.ca maps). TELUS and Bell run one shared radio network but split the building: ISED lists TELUS on the sites in the West and Québec, and Bell on those in Ontario and the East. Purple and blue together are the network a TELUS customer uses. Switch carriers to compare, filter to 5G, and zoom into a city to see which way each antenna faces. Switch to Who carries customers to see whose towers each carrier's customers use where it hasn't built its own.

The map shows whose towers carry TELUS customers. This is who pays for them. Bell and SaskTel build 44% of the network TELUS customers use. La Caisse owns half of TELUS's towers. Lenders hold $25.9B of net debt. Each one lowers what TELUS has to build or borrow, and each one takes something back.

TELUS fibre now reaches 3.8M homes and businesses in BC, Alberta and Eastern Québec. Past the last fibre and copper line, about 400,000 Canadian households still can't buy 50/10 internet, including 1 in 8 rural homes. Starlink has become a major Canadian ISP in that gap, with 500,000+ subscribers (about one for every six TELUS internet customers), and none of what they pay goes to a Canadian carrier. TELUS can take a share without building anything. Next to $7.4B of EBITDA the share is small. The capex it avoids and the customers it keeps are what matter.

Starlink reaches every corner of Canada from the sky. TELUS reaches 90% of Canadians with 5G but only part of the land. Where both reach, TELUS can win the home: 5G home internet, a hub for the truck, the farm's and the family's connected devices, and service a satellite company doesn't offer. Beyond TELUS's towers, Starlink is the only fast option, and TELUS should partner (section 03). Drag the map to compare.

Every dot is a block of subscribers in motion. Orbs are sized to each carrier's base and pulse as they gain and lose. Freedom keeps pulling from the incumbents; TELUS bleeds slowest. Tap an orb to isolate its flows.

Canada at true proportion — you can see the landmass each carrier has to cover. Every province is filled by its leader's colour, and brightness scales to subscriber mass, so the population reality cuts against the geography: Nunavut is vast and nearly empty, Ontario is compact and holds 14.81M. Click a province for the full breakdown.

One greater view: each province emits a live signal wave whose strength blends market dominance with the spectrum that carrier actually holds — more MHz = a taller, denser, faster wave that carries more energy. Pick a carrier tab to watch its air light up province by province with its spectrum footprint, or Overall to see every carrier stacked. Flip prepaid/postpaid. The leader's wave pulses brightest, and its MHz shows on the right.

Everything on one frame: every carrier's subscriber base stacked into postpaid (solid) and prepaid (hatched), sized to real scale. Rogers leads, then Bell, TELUS, Freedom — together ~86% of a 38.2M market. Bars animate up on the shared beat.

The gap between these two is the whole story of Canadian coverage: carriers cover a small slice of the landmass but nearly all the people, because Canadians cluster in cities.

Postpaid (monthly bill, credit check, contract) is ~90% of the base — low churn, high ARPU, the profit core. Prepaid (pay-ahead, no contract) is where budget brands and newcomers live. Each carrier plays both, but with different brands.

A telecom company is really two companies sharing a brand. Wireless sells mobility — air, towers, spectrum, a subscriber who can churn next month. Wireline sells a physical drop into a building — fibre in the ground, a 20-year asset, a household that almost never leaves. They have opposite economics, and TELUS wins differently in each.

The most important wireline matchup in Canada. Here's the honest read: both run the same underlying technology — XGS-PON, 100% fibre to the home, symmetrical speeds. Neither has a physics advantage. The war is won on footprint, execution, price certainty and who got there first. Hover a row.

Every call, stream and text follows this path in milliseconds. Packets flow outbound in amber, return in teal. Click a node to open its full technical spec below — hardware, protocol, latency budget and what breaks if it fails.

5G promises single-digit millisecond response. That budget gets spent across every hop — and it's why the CU and edge compute moved closer to the tower. Watch the waterfall fill.

Your data gets wrapped in envelopes on the way down and unwrapped on the way up. Hover a layer to see what it does and where it runs.

The tower used to hold everything. Modern networks split it apart and centralise the brains — cheaper to densify, and the reason 5G can slice itself into private networks. Click an architecture to watch the topology rearrange.

One physical trade-off governs the entire industry: low frequency travels far but carries little; high frequency carries enormous data but barely travels. The waves below are drawn to their real relative wavelength — watch how the low band stretches across the map while mmWave crowds into a few hundred metres.

Spectrum is licensed by ISED through auctions, then held as time-limited licences per geographic tier. Bars are scaled to total holdings — hover for the band split.

Telecom is a capital machine: buy spectrum, build network, rent it monthly, repeat. Enormous fixed costs, near-zero marginal cost per extra gigabyte.

Four phases. Listen before you move, commit before you're comfortable, and put the first visible win on the board before day 100.

The year you'd walk into, counting up in real time. Each bar fills toward its full-year mark; the thin line is how much of 2026 has gone. The right-hand column is what the second half has to deliver.

A new CEO gets one credibility budget. I'd spend it on three promises that are specific, measurable, and survivable.

Half of a first-100-days plan is the discipline to not act. These are deliberate holds — each one a mistake a new CEO is tempted into.

If I've done this right, here's what's true on day 100 — each item verifiable, no vanity metrics.

What the whole company's Q4 looks like if the 2026 trend holds, counting up from Oct 1. Each bar is last year's Q4; the solid part is Q4 2026 so far and the marker is where the projection lands.

Every square is a market or segment in Ontario and Quebec, coloured by who owns it today. Purple is ours. Squares outlined in teal are in play this quarter — where promos roll off, students and newcomers arrive, and Black Friday lands. Click any square to see the play.

Six years of Canadian wireless pricing in one chart. The government set a 25% reduction target off a 2020 benchmark; ISED confirmed it was hit by Q1 2022. Then the floor kept falling as flanker and discount brands went to war — and since late 2025 prices have been climbing back. In April 2026 Fido, Koodo and Virgin Plus all raised prices in near-lockstep.

Every carrier runs two price books. In-market is what you see advertised — the acquisition offer, public, competitive, measurable. BTL (below-the-line) is what the retention desk hands out when you threaten to leave: discretionary credits, typically $20–$60/month for 6–24 months, never advertised, never in the price list. The gap between them is where margin quietly leaks.

Acquisition spend modelled per channel — winback runs on owned data and brand equity (cheapest), in-market needs paid media plus a device-balance buyout (most expensive), flank runs self-serve on the flanker brands (cheap by design).

Named owner per workstream, three hard gates before anything ships, and the risks I'd want on record rather than discovered mid-quarter.

Q4 has a shape. Load the machine in October, spend hard through Black Friday and the holiday upgrade window, then convert volume into depth before the quarter closes.