Channel Diversification for Q4: Highlights from Nextdoor, Roku, and AppLovin
Why It Matters
Every year, as we approach Q4, it seems like the costs for advertising on platforms like Meta and Google keep climbing higher. The discussion around diversifying our spending always comes up, but not much really changes. This guide breaks down how to effectively allocate your holiday budget. It highlights exactly what each platform truly values, and it gives you the tools to make a convincing case internally, backed by solid numbers instead of just the usual vague arguments about upper-funnel performance.
Who This Is For
This message is for media buyers and growth teams looking to step beyond Meta and Google before the Black Friday rush. It's especially helpful for those who may have written off options like CTV, Nextdoor, or AppLovin as "too tricky to measure" without actually trying out the available attribution methods. Now is the perfect time to explore these alternatives!
Every Q4, the same conversation happens inside growth teams: Meta and Google are getting more expensive, we know we should be testing something else, and we probably won't get around to it before Black Friday.
I brought together three platforms making a case to prioritize those new-channel tests for a recent Foxwell Digital webinar. Gareth Domingo of AppLovin, Derek Chung of Nextdoor, and Benjamin Hardy of Roku were the stars of a great conversation about where incremental holiday dollars can actually go. The video is below, but if you want the ClifsNotes, here are the highlights.
Three very different arguments for attention
The panel opened on what each platform is actually selling. The answers were less alike than expected.
AppLovin is selling forced attention. Gareth described in-game ad units as full-screen takeovers with no content above or below — 100% share of voice, unskippable for a stretch, and averaging around 35 seconds per impression. He argued that the audience is a valuable pool of shoppers that consumer brands largely aren't reaching yet. Critically, the platform doesn't optimize toward reach or traffic: "Everything we optimize towards is actual purchases."
Nextdoor is selling verified identity. Every user goes through address verification before they can post or engage, and Derek noted that roughly one in three U.S. households are on the platform. That produces neighborhood-level targeting no one else can match, plus inferred signals (e.g. homeownership correlating to income,) that build an audience makeup Derek framed as genuinely incremental, since much of it doesn't exist on other platforms.
Roku is selling the biggest screen in the house. Benjamin came out of the gates with an important point: not all impressions are created equal. CTV delivers completion rates in the 98–99% range, audio on, alongside professionally produced content rather than a UGC feed — what he called the "Look mom, I'm on TV," factor for brand legitimacy. And because television is a household environment, you're often reaching two to three people per impression, effectively halving your CPM out of the gate.
How to make the case internally
I asked the panel to help viewers get a smart approach a common challenge: how does a media buyer sell a less familiar, harder-to-measure channel to an executive?
Benjamin pushed back on the premise. He'd take "harder to measure" off the table entirely — the upper-funnel, awareness-only assumptions inherited from linear TV don't apply to CTV. Roku sits at roughly 100 million households with about half of all CTV time spent on its devices, all logged-in registered users, which supports an identity graph that matches against whatever conversion events you send. He also made the incrementalist's pitch: you don't need to enter a channel all at once. Take a proven tactic (e.g. site-visitor retargeting excluding purchasers) and ask whether a second Instagram impression is worth more than one in-feed impression plus one on television. On the cost objection, he cited a Q2 average CPM around $7.50 and noted Roku offers new advertisers a 1:1 match credit on a first campaign.
Derek's angle was reframing rather than replacing. Nextdoor isn't competing for the social budget so much as offering "a relatively low-risk incrementality test" against households a brand is currently underreaching. He emphasized how little friction is involved: the pixel fires the same standard events as Meta's, integrations work the same way, and existing creative, images, and descriptions port over directly. His team has seen advertisers up and running in under an hour.
Gareth’s argument was built on scale. AppLovin runs over $11 billion in annual ad spend on a pure performance basis, so the channel doesn't cap out at a few thousand dollars a day. On measurement, he made a point of noting the web conversion business launched with click-through attribution only — deliberately the most deterministic option, when most platforms angle for view-through credit. Advertisers have brought MTA tools, MMM, incrementality tests, and geo holdouts, and compared AppLovin against their best existing channels. His summary of what advertisers appreciate: no alternative attribution methodology required to justify the spend.
Real numbers, real platform breakdowns, no vague "diversify your channels" advice.
What the brands winning on these platforms do differently
Asked what separates successful advertisers, the panel converged on a theme: familiarity gets you in the door, adaptation gets you results.
Gareth's example was a small Israeli cookware company doing around $4 million in revenue when it joined. Rather than treating the channel as a side test, the team asked a first-principles question — if we couldn't advertise anywhere else, how would we make this work? They studied the in-game ad experience and rebuilt their funnel and messaging to be native to it. Within about 12 months they went from $4 million to $60 million, and are now on a run rate toward $80 million.
The tactical translation matters. Social teaches short-form, organic-feeling, entertaining creative. AppLovin rewards the opposite: 45- to 60-second videos with narrative depth and direct-response structure — problem, agitation, escalation, solution, offer. Gareth's shorthand was a Billy Mays infomercial. Landing pages need rethinking too, since the channel sends more curious, top-of-funnel traffic; long-form sales pages and quiz mechanics are performing well. And all of it demands creative volume, because any engagement-based algorithm is hungry for concepts and variations.
Derek pointed to advertisers leaning into Nextdoor's proprietary data layers — local substitution that names a specific neighborhood rather than a metro, multi-location ads that serve a storefront-specific creative to the neighborhood that will visit it, and a weather targeting beta that only serves your puffy-jacket ad once temperatures drop at the user's verified address.
Benjamin added two important “don’ts.” First, don't loosen your standards for a new channel — bring your geolift and incrementality tests on day one, and ask why if a partner resists. Second, don't evaluate channels in a vacuum. It's common for CTV CPA to come in above social, while a geotest shows blended CAC improving materially in the markets where CTV is running.
Why Q4 specifically
Roku: supply isn't the question on television; demand is. Benjamin expects to see thousands of Black Friday offers in his standard feeds and comparatively few on TV — an arbitrage between where attention sits and where advertisers have shown up.
Nextdoor: neighbors outspend the general U.S. consumer during gifting season. Awareness and consideration campaigns beat their full-year averages in Q4, and promotional, offer-led creative outperforms community-style creative. The new-mover engine has been especially strong for home security, home services, and consumer electronics.
AppLovin: mobile gamers are mainstream consumers, skewing 25–55 with disposable income and household decision-making power. Unskippable impressions mean 100% of them see your offer, and scaling is fast — brands can read performance in four or five days and step from $1,000 to $50,000 a day with limited impact to CPA or ROAS.
Before we get into the Q&A and promo offers, here’s the full video if you’d rather dive deep:
From the Q&A
On frequency. All three default to algorithmic control. Benjamin declined to name a magic number, noting categories respond differently, but said Roku typically sees lower frequency than advertisers expect — often one to two per week on broad audiences — and suggested adding a re-engagement campaign or tightening audiences if you suspect higher frequency would help. Derek recommends starting around 10 a week on Nextdoor, since most neighbors are weekly rather than daily actives, then opening or tightening from there. On AppLovin, per Gareth, the average purchaser sees eight unique ads across ten impressions.
On creative support. All three offer it. Nextdoor has a dedicated account management and ad strategy team with relatively low spend minimums, plus trained general support. Roku provides creative services at no cost, including third-party partners who can custom shoot or use AI to convert vertical assets into big-screen horizontal, with in-UI enhancement tools in development.
Closing beefs
I closed with the segment we run on the Scalability School podcast — the thing each guest wants to shout from the rooftop.
Gareth reached back to September–October 2024, when the product first got real traction without a self-serve platform and the team onboarded advertisers manually around the clock. He compared it to the Facebook ads moment of 2011–12: enormous white space, minimal competition. AppLovin now reaches over a billion people a day with thousands of advertisers, against platforms with millions. His read: bottom of the second. If you tried it a year or two ago and it didn't work, try it again.
Benjamin argued that CTV has stopped being out of reach. It overtook linear TV last year and supply will only grow. He made the case for buying direct rather than through a DSP — same inventory, fewer fees along the way — and for scaling gradually enough to prove results to yourself and your team.
Derek stressed context. Nextdoor users aren't doomscrolling for entertainment; they're there for utility, asking neighbors for help and recommendations. Showing up alongside those conversations, among verified users, earns a willingness to take a chance that other environments don't.
Promo offers
Attendees got first crack at special offers from our panelists:
Nextdoor: Spend $5k, get $5k in ad credits.
Roku: Spend $5k, get $5k in ad credits.
AppLovin: Spend $5k and get $5k in ad credits + an extra $1k if you launch within 24 hours of creating an account.
Want an intro to any of these panelists? Hit me up at andrew@foxwelldigital.com.
Three new channels, three different playbooks and Founders members are already testing all of them. Join the Foxwell Founders Membership to compare notes with operators actually running the spend.
