Information and cultural industries

A community paper, a trade magazine or a small book list is one of the cheapest businesses in this research to start: no plant, no licence, and 1,420 of Canada's 1,972 publishing establishments employ fewer than ten people [A]. The cut is what has happened to the revenue the model rests on. Statistics Canada puts newspaper publishers' operating revenue at $1.6B in 2024, down 17.9% in two years [A]. Print advertising fell 34.3%, which is expected. The finding that matters is that digital advertising fell too — down 11.9% to $315.1M [A]. The standing argument for a new publisher was that the reader could be found again online and the advertising would follow; the agency's own series says the advertising went to the platforms instead, and publishers are now shrinking on both sides of the ledger. The industry held a 3.2% operating margin only by cutting expenses 19.1% [A] — it is being managed down, not rebuilt. Postmedia shows the same thing at scale: fiscal 2025 revenue of $431.5M rose only because it bought Saltwire, advertising grew 2.7% without it, and the year closed on a $77.3M net loss [A]. Periodicals are steadier — $1.0B in 2023, up 1.9% over two years — but expenses grew 5.2% and the margin fell to 6.3% [A]. What survives inside this group is publishing as a by-product of something else: an association's magazine, an events business with a title attached (events, conferences and trade shows were 5.8% of sales). That is a reason to own an audience, not a reason to enter publishing. Book and directory publishing were not examined. The editorial and rights software sold to publishers is screened separately at 513.

This is the code software companies themselves file under, and by the numbers it is the most attractive group in its sector. Statistics Canada reports software publishers at $27.9B of operating revenue in 2024, up 15.6%, with 68.8% of sales to clients outside Canada [A]. No plant, no licence, a global market from the first day, and 1,076 of the 1,808 establishments employ fewer than ten people [A]. The screen does not find a clean cut at this level, and says so. The nearest thing to one is in the same release: on $27.9B of revenue the industry carried $26.6B of expenses — an operating margin of 4.7% [A], with salaries the largest cost. An industry growing at 15% and keeping under five cents on the dollar is one where the product is cheap to make and the customer is expensive to reach; growth is being bought with payroll. That points at distribution as the binding constraint, and it is the factor recorded here. But distribution is not a property of 'software publishing'. It is a property of the buyer: selling to dental practices, to carriers and to game players are three unrelated problems with different incumbents, sales cycles and prices, and an average over them has no referent. That is why this research screens software by the industry it is sold into — the /software lens holds those records — rather than as one market here. A full study at this level would have nothing to test; the test belongs to each vertical. Video game publishing (513212) has hit-driven economics of its own and was not examined.

A small-market radio station is the enterable thing in this group: a CRTC licence, a transmitter, a few staff, and the merchants of one town as advertisers. 673 of Canada's 1,189 broadcasting establishments employ fewer than ten people [A], and stations change hands regularly. The pre-screen cut this on licences and spectrum. The screen finds the licence is the smaller problem. The revenue the licence protects is shrinking. Trade reporting of the CRTC's 2024–25 market report puts commercial radio at $1.06B, down 2.6%, and private conventional television at $1.17B, down 7.3%, on a PBIT margin of −40.4%, while online services took about 40% of all broadcasting revenue [B]. A licence is a protected position in a market advertisers are leaving, and the protection does not extend to the platforms taking the money. Stingray shows what it takes to stand still: its radio segment held $132.4M of revenue, flat, at a 31.3% adjusted EBITDA margin in fiscal 2026 [A], with growth in digital advertising only offsetting the fall in airtime sales. That margin is a group result — shared programming, national sales representation and engineering spread across a portfolio of stations. The single station has the same decline without the shared cost base, and its eventual exit is a sale to one of those groups at the buyer's price. Television is cut more simply: the sector's operating margin is deeply negative before an entrant has bought a camera. Streaming and online audio have no licence barrier and sit outside this group.

A music streaming service is a licence business: every service rents the same catalogue from the same few rights holders, on terms set as a share of its own revenue. That makes the gross margin the thing an entrant is buying, and the incumbents publish it. Spotify took €17,186M of revenue in 2025 and spent €11,690M on cost of revenue, which it says consists predominantly of royalty and distribution costs — a 32% gross margin at 290 million paying subscribers [A]. Major-label content was about 72% of the label-delivered streams it served [A], so the four licensors that matter (Universal, Sony, Warner and the indie agency Merlin) negotiate with Spotify, Apple, Amazon and Google before they negotiate with anyone else. Below Spotify the margin thins out: SiriusXM's Pandora and Off-platform segment paid $1,308M of revenue share and royalties on $2,141M of flat revenue, while Pandora's monthly active users fell 5% to 41.1 million [A]; Deezer, with eighteen years of operating history, reached its first positive adjusted EBITDA in 2025 — €9.7M on €534M of revenue, under 2% [B]. The statutory side is no cheaper for a newcomer: in the US the Copyright Royalty Board's Phonorecords IV schedule raises the songwriters' headline rate on interactive streaming from 15.1% of service revenue in 2023 to 15.35% in 2027 [B], and non-interactive streams pay SoundExchange per performance — $0.0028 per performance in 2026 rising to $0.0032 in 2030 for commercial broadcasters' streams, with a per-station minimum [A]. None of these rates falls with volume in a way a small service can exploit; scale is what earns the direct deals and the marketing reach, and three of the largest competitors (Apple, Amazon, Google) run music inside a bundle where it does not need to make money at all. Defensibility decides it: an entrant pays the same royalty share as Spotify, offers the same catalogue, and has no product it can withhold from the incumbents. The niches that survive do so by owning a rights or audience corner the generalists ignore — DJ mixes (Mixcloud spent years licensing through collecting societies before signing a direct, multi-year deal with Universal in 2018 [B]), DJ-tool streaming (Beatport absorbed Beatsource into one service [A]), hi-res and classical (Qobuz, IDAGIO), and emerging-market and hip-hop catalogues (Audiomack) — and the best-known independent among them, SoundCloud, was forecast at only €2M of positive EBITDA on €288M of revenue for 2023 when its owners explored a sale above $1B [B]. Canada adds a moving regulatory cost: the CRTC's 2024 Online Streaming Act decision required large unaffiliated services to pay 5% of Canadian revenue into content funds; payments were stayed from December 2024, Spotify was among the challengers, and a July 2026 Justice Department letter to the Federal Court of Appeal said the government intends to eliminate the base contribution [B]. Performance and neighbouring-rights royalties still flow through SOCAN and Re:Sound. This record differs from its neighbours: 5161 is a licensed over-the-air radio station selling local airtime; 5122 is the hosting and ad-insertion software podcasters use; 513 is title and rights software sold to publishers. This is the consumer-facing service that licenses recorded music and streams it.

Inside a group dominated by national carriers there is a real small-operator business: the regional provider that builds its own fibre or fixed-wireless network in places the majors have not reached. The counts show it — 1,034 of Canada's 1,777 carrier establishments employ fewer than ten people, against 38 that employ 500 or more [A]. The demand is not in question; connectivity is the one utility households will not give up. The cut is what must be sunk before the first bill goes out, and who the entrant is out-building. BCE spent $3,700M of capital in 2025 — 15.1% of revenue — and that was a reduced year, on operating revenue that grew 0.2% [A]. A network business at maturity reinvests about fifteen cents of every revenue dollar simply to stay current, and keeps a 43.6% adjusted EBITDA margin [A] only because the network underneath was paid for over decades. The regional entrant faces the same cost per home passed with none of that base, in territory that was left unbuilt precisely because density would not carry the cost. That is why rural builds lean on public broadband funding — and a subsidised build has the funder's timetable, coverage obligations and reporting attached. The exposure that does not go away: once the entrant proves a territory pays, the incumbent can overbuild it out of a capital budget larger than the entrant's enterprise value, or a low-earth-orbit service can take the most remote customers with no build at all. Resale over someone else's network is a different proposition and is screened at 5179; billing and operations software for carriers is screened separately at 517.

Nobody with ordinary resources launches satellites, and the pre-screen cut this group on capital. But the definition also covers firms distributing the services of other satellite operators, and that is what most of the group is: 79 Canadian establishments, 55 of them with fewer than ten people [A] — integrators putting terminals on mine sites, vessels, northern communities and backup links for enterprise networks. That business needs a van and a dealer agreement, not a constellation. The cut is that the thing being resold is losing its price, and the new supplier does not need a reseller. Telesat, Canada's operator, reported 2025 revenue of $418M, down 27%, and adjusted EBITDA of $213M, down 45% [A], as demand for geostationary capacity fell away. Its answer is to become a low-earth-orbit operator itself: $708M of capital expenditure in 2025 and a further $1.0–1.2B planned for 2026 on Lightspeed [A] — a 2026 commitment of more than twice the $418M the whole company earned in 2025. The reseller sits beneath that fight. Its historic margin came from the scarcity of capacity and the difficulty of installing and pointing a terminal; low-earth-orbit services ship a self-installing terminal and sell to the end customer directly at a published price. What is left is integration labour — mounting, networking, managed service for an industrial site — a small field-services business whose supplier is also its competitor, not a telecommunications margin. Billing and operations software for this sector is screened separately at 517.

The pre-screen called this a genuine small-operator entry, and on the counts it is the most populated corner of telecom: 3,019 establishments, 2,554 of them with fewer than ten employees and none above 499 [A]. A reseller needs no network — it buys regulated wholesale access from the telephone or cable company, puts its own brand and support on top, and sells home internet or business voice. The screen tested that and the pre-screen does not survive. The reseller's largest cost is a tariff paid to the company it competes with at retail, and the record of the last five years is what happens under that arrangement. Reporting of the CRTC's 2025 market report puts independent wholesale-based providers' share of home internet at 8.4% in 2020, 8.0% in 2021, 6.1% in 2022, 5.0% in 2023 and 4.2% in 2024 — halved in four years [B]. Over the same period the larger independents stopped being independent: Bell bought EBOX and Distributel in 2022, and Quebecor, Videotron's parent, bought VMedia the same year [B]. The incumbent can discount its own flanker brand below the wholesale rate plus the reseller's costs, contest every rate decision for years, and then buy whoever is left; the reseller can do none of those things back. Hosted business voice is the same shape with a different supplier — the platform behind a white-label VoIP service also sells direct. An incumbent that sets your input price, competes for your customer and is the natural buyer of your business is not vulnerable. The 'all other telecommunications' tail of this group was not examined. Billing software for the sector is screened separately at 517.

The demand is the least ambiguous on the entire list, and it is the cause of the opportunity the grid study at 221121 examines. The problem is what it costs to serve. Amazon Web Services earned US$128.7B in 2025, from US$107.6B in 2024, inside a company that spent US$131.8B on property and equipment in the same year [A] — a capital programme larger than the segment's whole revenue. Equinix is the closer comparable, because colocation is what a non-hyperscaler could plausibly sell, and it tells the same story: $9.217B of 2025 revenue on roughly $4.3B of capital expenditure, about forty-seven cents of capex for every revenue dollar, with $3.655–4.155B guided again for 2026 [A]. In Canada the unit of entry is now quoted the same way — Innovation, Science and Economic Development Canada describes Bell's Saskatchewan AI hub as up to $52.5B of capital investment for up to 900 MW [A]. Against that, the 1,163 Canadian establishments in this code are small: 538 employ fewer than five people [A]. They are not building capacity, they are managed-hosting, web-hosting and data-processing firms reselling someone else's. An entrant cannot buy its way to a cost position here, because the cost position is set by counterparties spending more in one year than an entrant could raise in a lifetime, and the scarce input — interconnected grid power — is allocated years ahead to whoever is already in the queue. What is left enterable sits on top of that infrastructure: sovereign or regulated hosting, managed migration, specialised workloads. That is a labour business wearing an infrastructure name. The grid study at 221121 is how a small operator sells into this wave without building it.

Two unrelated things share this code, and the record has to take them apart before it can cut either. Libraries and archives are public institutions, not a market to enter. They are funded by municipalities, provinces and universities, and nobody opens a competing one. The establishment counts show it: Saskatchewan has 284 of Canada's 1,520 establishments, more than Alberta or British Columbia [A] — not an industry cluster but, most plausibly, a province-wide branch system counted branch by branch. What is sold to libraries is screened separately at 519. The enterable half is the web portal or online information service: a niche directory, a reference site, a data service supported by advertising or subscriptions. The cost to start is close to nothing, which is the problem. Such a site has no distribution of its own — its readers arrive through a search engine — and the search engine is both the gatekeeper and the largest seller of the same advertising. Alphabet's Google Search & other revenue was US$224.5B in 2025, up from US$198.1B [A]; that is the scale of the counterparty that decides whether a portal's pages are seen, and that increasingly answers the query on its own results page instead of sending the visit. A portal that ranks has a business at the search engine's discretion, earning an advertising rate set in the search engine's auction. The versions that survive own their audience directly — a login, a subscription, data nobody else has — and that is a different business from the one this code describes. Archives and the 'all other information services' tail were not examined.