Every one of these four levers has, at some point this year, been covered here as its own individual story. Put together, side by side, they're actually a single budget-allocation problem, and most link building programs are still funding all four of them as if they were fully interchangeable line items on the same spreadsheet.
Four levers, one budget
Earned media is the biggest, most consistent driver by the numbers Something Inc. has already published: Muck Rack's analysis puts 82% of tracked AI citations behind earned media coverage, meaning a brand mentioned or quoted in genuine third-party journalism. It's also the most expensive and the slowest to build, requiring real relationships, real news hooks, and real subject-matter authority a reporter has a reason to call on. A single well-placed feature in a trade publication a target buyer actually reads can outperform months of lower-tier placements, but that kind of placement is earned through sustained credibility, not purchased through a single pitch, which is exactly why it doesn't show up on a quarter's budget spreadsheet as a predictable, schedulable line item the way a paid program does.
Preferred Sources is the opposite profile: fast, cheap relative to earned media, and entirely controlled by Google. The program has expanded into AI Mode and AI Overviews, curating 345,000-plus sources and delivering roughly double the click rate of a standard citation once a domain is included. Getting in is a light-touch inclusion process, not a relationship-building campaign, which is exactly why it's a lower-durability bet: it's Google's list, subject to Google's criteria, and a brand has essentially no negotiating leverage over its own inclusion. There's also no visible appeals process if a domain gets dropped, and no contractual guarantee the program continues to exist in its current form at all, the same structural risk every Google-owned curation layer carries regardless of how generous the terms look today.
Review platform management is the one with a genuinely sharp threshold effect, and the one most link building budgets still under-fund relative to how much it actually moves. Seer Interactive's Trustpilot-commissioned study, covered here in detail, found brands with no review profile sit at a 1% median AI citation rate. Adding even a minimal profile, 1 to 13 reviews, jumps that to 53.5%. An actively managed profile with around 81 median reviews reaches roughly 81%. That's the steepest, most controllable curve of the four, and notably, 99.5% of the citations in that study arrived because the review page itself ranked well in organic search and got pulled into an AI answer from there, not because an AI engine queried the review platform directly. That means a brand's own review-page SEO is doing real work inside a channel that looks, at first glance, entirely platform-owned.
The fourth lever is the newest and least proven: the UK CMA's order, covered in detail elsewhere this week, requiring Google to cite content with proper attribution and a clear link whenever its AI features use it. If that requirement holds and expands globally the way Google's own opt-out feature did, it changes the value of every citation a brand already earns, since a citation with a guaranteed visible link is worth more traffic than one without. It's not a channel to invest in directly, since there's nothing to build toward yet, but it's a multiplier that could apply to everything else on this list once it ships, and a program that's already strong on the other three levers is the one best positioned to benefit the moment it does.
The decision table
| CHANNEL | CITATION SHARE / LIFT | WHO CONTROLS IT | DURABILITY |
|---|---|---|---|
| Earned media | 82% of tracked citations (Muck Rack) | Brand builds relationships and authority; no single platform gatekeeps it | High: survives any single platform's algorithm or policy change |
| Preferred Sources | ~2x click rate once included | Google decides inclusion criteria and the list itself | Low: one policy change removes the advantage entirely |
| Review platform management | 1% to 81% citation rate across profile tiers | Brand manages review volume and its own review-page SEO | Medium-high: depends partly on one platform's continued relevance, but the SEO layer is brand-owned |
| UK CMA attribution mandate | Not yet measurable; multiplier on existing citations | Regulator-driven, not brand-controlled, but non-optional for Google if enforced | Unknown: depends on enforcement and global rollout |
Why durability should outrank size
It's tempting to read the 82% Muck Rack figure and the 345,000-source Preferred Sources number and conclude the second is simply a faster way to punch above your weight while the earned-media relationships compound in the background. That's a reasonable tactical read, and Preferred Sources is genuinely worth pursuing as a light-touch addition. But it shouldn't be budgeted as a substitute for earned media, because the two have almost opposite risk profiles. Earned media citations are the byproduct of relationships and authority a brand actually owns; Google can change how it weights or displays a Preferred Source entry without a brand having any say in it at all.
Something Inc. has already watched this exact same substitution mistake play out before, just with different channel names attached. Teams that leaned hard into Reddit as a citation channel earlier this year, back when Reddit's share of ChatGPT citations was still climbing, got a hard lesson in platform-dependency risk when that share collapsed within weeks after a licensing dispute changed what Google's crawlers could access. Nothing about the brand's own content or authority changed in that collapse. The channel underneath it changed, unilaterally, and every program that had treated Reddit presence as a durable citation strategy rather than a single channel among several took the full hit at once. Preferred Sources carries the same structural exposure, just with a different company holding the lever this time.
That history is exactly why the durability column in the table above should carry real weight in a budget conversation, not just the citation-share column. A channel that delivers a smaller number today but doesn't depend on one company's ongoing goodwill is worth more over a two- or three-year planning horizon than a channel delivering a bigger number that could disappear with a single policy update nobody outside Google gets advance warning about.
This isn't a hypothetical risk specific to AI search, either. Foundation's analysis of 5.1 million AI responses already found third-party citation, not owned content, drives the overwhelming majority of B2B AI recommendations, and every third-party channel on this list carries some version of platform dependency risk. The question isn't whether to accept that risk, since there's no channel on this list, including your own owned content, that's fully insulated from a platform decision somewhere upstream. It's how much of the total budget should sit in the channels where the brand at least retains some agency versus the ones where a single external party holds all the leverage.
How to actually split the budget
A reasonable starting allocation, given the durability spread in the table above, weights earned media as the anchor, somewhere around half of total link building and digital PR spend, since it's both the largest measured driver by far and, structurally, the single most durable channel on the list. Review platform management deserves a meaningfully larger allocation than most programs currently give it, given how steep and controllable that citation-rate curve is, and given that a large share of its value flows through a lever, the review page's own organic ranking, that a link building team already knows how to work.
Preferred Sources belongs in the budget, but as a maintenance-level line item, not a strategic bet: confirm inclusion criteria are met, keep technical access clean, and don't build a program's core citation strategy around a list a single company controls unilaterally. The CMA attribution mandate doesn't get its own budget line yet, since there's nothing concrete to build toward, but it belongs on the same quarterly watch-list as the opt-out audit itself, since a shift in how much a citation is actually worth changes the relative value of everything above it in this table. Treat it as a reason to double down on the durable channels now, not a reason to wait and see what happens before committing budget anywhere at all.
Run this allocation as a genuine quarterly review, not a set-and-forget split. Every channel in this table has moved meaningfully in the past two months alone, Preferred Sources expanding its footprint, the review-profile data getting more granular, and an entirely new regulatory lever appearing that didn't exist when most 2026 link building budgets were first set. A B2B SaaS team that locked in an allocation in January is very likely underweighting review-platform management and hasn't yet accounted for the CMA order at all. Rebuilding this table every quarter, with current numbers, is cheaper than discovering the gap the hard way when one channel's citation share moves and the budget hasn't kept up.
The teams most likely to get this wrong are the ones treating link building budget as a single line item reported up as one number: dollars spent, links acquired, citation share moved. That reporting structure hides exactly the risk this table is built to surface, because it can't distinguish a program that's 70% dependent on a single platform's continued goodwill from one that's genuinely diversified across four structurally different levers. Break the reporting apart by channel before the budget conversation happens, not after, so the durability question gets asked while there's still time to shift dollars rather than after a channel has already moved and the quarter's spend is locked in.
One more practical note on sequencing: review-platform management is the fastest of the four to show measurable movement, often inside a single quarter, since it doesn't require building new relationships from nothing the way earned media does. Teams looking for a near-term proof point to justify reallocating budget away from a Preferred-Sources-heavy mix should start there, build the review-page SEO layer properly, track the citation-rate lift directly against Seer's published tier benchmarks, and use that concrete, internally-generated result to make the actual case for shifting more of next quarter's budget toward the durable side of this table.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Josh leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.