Think of AI as a turbocharged sketchpad and humans as the creative studio steering the brush. When algorithms crank out ideas at scale, people supply context, taste, brand memory, and moral sense. The best teams treat machine output like raw footage: useful, messy, full of flashes of gold that only a human can polish into an ad that moves people. That partnership is not a cage match; it is a relay, fast and strategic.
Start every campaign with a human brief that AI can actually use: clear objectives, audience archetypes, must have lines, and things to avoid. Let AI generate ten fast concepts, then run a human triage to pick the top two for expansion. Humans refine voice, emotional beats, and cultural nuance; AI scales variations and simulates channels. This workflow speeds experimentation and keeps brand intent intact while multiplying creative options.
Use constraints to get better AI results: restrict vocabulary, set tempo, demand a literal hook in the first three seconds for video work, or require a surprising twist in copy. Combine that with adversarial review where people seek off brand outputs, legal flags, and insensitive angles that models might miss. Keep a living style guide and short prompt templates so every teammate can recreate winning prompts consistently.
Measure human plus AI work with both hard and soft KPIs: conversion lifts, engagement patterns, creative decay, and qualitative audience feedback. Rotate human reviewers each campaign to avoid groupthink and invest in quick micro tests before full spends. When teams follow these steps, AI amplifies human judgement instead of replacing it, producing ads that are faster, cheaper, and more likely to actually be enjoyed.
Cookies once felt like magical crumbs that told marketers everything. With those third party crumbs baked away, brands cannot rely on spray and pray retargeting. The good news is that owning customer signals is both tastier and more defensible, and it reduces legal risk when data flows through your own gates. First party signals improve personalization, tighten attribution, and often boost margins.
Start by redesigning the value exchange. Offer exclusive utilities such as faster checkout, meaningful recommendations, and loyalty points in return for direct information. Use short preference centers, topic tags, and micro surveys to capture zero party intent. Gamify capture where sensible and make consent simple, transparent, and reversible; that builds trust and meaningfully raises data quality. Small UX wins compound quickly.
Back capture with solid plumbing: server side tracking, conversion APIs, identity resolution, and cleanroom partnerships with publishers or cloud providers. Push hashed segments to platforms instead of raw PII and run cohort level lift tests rather than relying on impressions alone. Pair those systems with a testing cadence so creative and media learn from real behavior, not assumptions.
If this seems like a lot, split it into experiments: optimize one onboarding field, test two incentive offers, and run a ninety day cohort lift. Treat first party data as a relationship, not inventory, and you will see the advantage compound. The cookie may have crumbled, but the brands that bake better experiences will be first to the dessert table.
Treat connected TV as a performance channel, not cable refinement. The living room screen is where cinematic storytelling meets traceable outcomes: impressions that can translate to site visits, app installs, and measurable lifts. That mix of scale and signal turns brand cinema into data-driven campaigns, so planning feels more like analytics and less like nostalgia.
Measurement has caught up. Deterministic IDs, household graphing, and cross-device attribution let advertisers connect a 30-second spot to downstream actions. Programmatic pipes allow real-time bidding on audiences, frequency caps keep waste low, and viewability plus conversion windows create clear KPIs. In short, performance metrics live natively on the big screen.
Creative must evolve too. Think shorter hooks, visual first frames, and built-in call to actions that work with a remote or companion device. Test variations rapidly, favor dynamic creative that personalizes offers, and use sequential messaging to move viewers along the funnel. A beautifully shot ad that does not drive action is lovely but incomplete.
Targeting and efficiency are the secret sauce. Household and cohort targeting beat blunt demos, retargeting reengages high intent viewers, and daypart optimization squeezes more value from budgets. CPMs can be premium for prime content but true cost per acquisition often drops because conversions are measurable. Treat reach plus conversion as the primary objective.
Start with a small experiment: pick one KPI, run a controlled lift test, and use a partner that provides attribution transparency. Scale winning creatives, keep creative cycles tight, and budget for measurement. If ads are judged by outcomes, connected TV is not a nostalgic upgrade but the modern performance channel you can prove.
Remember when an ad popping up mid-scroll actually stopped you? Those interruptions are so 2007. Our fact-checking kept coming back to the same truth: context and creator-led moments win because they trade interruption for resonance. Ads that feel native to the feed, timed for a mood or a micro-moment, turn passive skimming into real engagement.
Make creators collaborators, not read-from-a-brief mannequins. Give a tight objective, a clear hook window (first 2 seconds), and permission to remix the brand voice in platform-native language. Prioritize formats that reward rewatching and interaction — loopable edits, authentic reveals, conversational CTAs — so the content earns attention instead of demanding it.
Swap a slice of CPM mindset for experiments that track what actually matters: watch-through, saves, shares, comments and lift in brand affinity. Run creator-versus-brand A/Bs with consistent placement; you'll often find the creator-native version drives deeper metrics even when reach is smaller. The prediction held up: authenticity plus context beats blunt reach for durable ROI.
Ship a tiny playbook and iterate fast: Map context: identify platform micro-moments; Partner: recruit creators who own that moment; Format-first: design for native behaviors; Measure: favor engagement and lift over raw impressions. Small process shifts, big feelings — and the data keeps the receipts.
Stop reporting applause. When the goal is to move revenue, likes and views are just warm fuzzies. Replace vanity with value by asking what action truly nudges purchase behavior. That shifts the conversation from surface engagement to things that impact margin, inventory turnover, and repeat purchase. The measurement goal becomes business decisions, not ego.
Track outcomes that map back to cash: conversion rate, AOV, CAC, LTV, and ROAS but do not stop there. Layer in incrementality tests and holdout groups to see what would not have happened without the ad. Use experiment results to prune wasted spend, not to justify more of the same creative that only generated attention.
Start small and instrument everything. Tag value for every conversion, even micro conversions like email signups that predict future spend. Align attribution windows with purchase cycles, and collapse redundant metrics so teams focus on one north star. Tie creative tests to incremental revenue per dollar, then scale winners and kill losers fast.
Make dashboards that CFOs can read without a translator. Replace applause metrics with the numbers that change budgets and hiring. If a metric does not alter a decision about product, pricing, or distribution, archive it. Vanity metrics belong in bathrooms, not dashboards. Measure what moves money and watch strategy follow the data.
Aleksandr Dolgopolov, 07 January 2026