Video marketing in 2026: YouTube, Shorts, Reels, TikTok, CTV and live video strategy

Video marketing in 2026 means making short vertical clips for feeds, longer episodes for the television in the living room, and streaming ads that behave like performance media. US digital video ad buying passes US$80 billion this year, and YouTube alone now takes 13.8% of all American TV watch time.

What video marketing covers in 2026

Video marketing is the use of moving image to reach, persuade and serve customers. It splits into four jobs with different skills and budgets: short vertical video for recommendation feeds, longer horizontal video for the television screen and for search, paid video bought across YouTube, social and streaming apps, and functional video on your own site doing sales work (demos, testimonials, onboarding).

Treat them separately because they fail for different reasons. A brand film that flops on connected TV usually failed on targeting or frequency. A Reel that flops failed in the first second. And a product demo nobody plays usually failed because it was 11 minutes long and sat below the fold.

The money has moved. The IAB’s 2026 Digital Video Ad Spend and Strategy Report, published 5 May 2026, projects US digital video ad spend above US$80 billion this year, growing 11% and taking more than 60% of total TV and video spend for the first time. That is not a forecast. It describes budgets already committed.

Where the viewing actually is

Plan against measured attention, not against the platform you personally use. Nielsen’s May 2026 Gauge, released on 28 July 2026, recorded streaming at 48.6% of US television watch time, a record, with cable at 20.4% and broadcast at 19.2%. YouTube was the single largest distributor at 13.8%, its third consecutive month in front, and up 13% year on year according to MediaPost’s read of the same data.

Notice what that does to old planning habits. YouTube is not a “digital extra” bolted onto a TV plan. It is bigger on the TV set than Disney or Netflix. Amazon Prime Video and Peacock both grew 32% year on year, helped by live sport.

YouTube: long-form versus Shorts

YouTube carries both halves of the video business inside one app, and marketers keep treating it as one thing. Alphabet reported US$11.1 billion in YouTube advertising revenue for Q2 2026, up about 13% year on year, roughly 13.5% of Google’s total ad business. Sundar Pichai used the earnings call to point at the living room rather than the phone.

“More than 1.7 billion unique viewers watched videos connected to the 2026 FIFA World Cup during the tournament.” Sundar Pichai, Alphabet Q2 2026 earnings call

The practical split works like this. Shorts are a discovery engine: cheap to make, huge reach, weak intent, and viewers who mostly do not remember which channel they came from. Long-form is where subscription, trust and search demand live, and it is the format that plays on a television. A channel that only makes Shorts builds view counts. A channel that only makes long-form grows slowly. The pairing that works is Shorts as the top of the funnel feeding a long-form library that answers real questions.

On the ad side, YouTube spent late 2025 and 2026 making Shorts inventory behave more like the rest of the platform. In December 2025 it enabled comments on Shorts ads, allowed creators to add website links inside branded Shorts, and extended Shorts ad placements to mobile web and TV surfaces. YouTube cited Kantar research putting an 8.8% lift in purchase intent for creator ads on Shorts. Treat vendor-supplied lift numbers as directional, not as proof.

TikTok, Reels and the feed video layer

Short vertical video is now the default unit of social. It is also the format marketers rate most highly: HubSpot’s 2026 State of Marketing data shows short-form video at 49% for ROI, ahead of long-form video at 29% and live streaming at 25%, and 93% of marketers calling video important to their strategy. Only 37% plan to increase video investment in 2026, which suggests most teams think they are already spending enough.

Creative rules are close enough across TikTok, Reels and Shorts that one shoot serves all three. Distribution rules are not. TikTok pushes hardest to people who do not follow you, Reels leans on shares and sends, and Shorts increasingly plays on a television where a phone-shot vertical clip looks poor. Post natively; watermarked reposts still lose reach.

Governance matters more than it used to. TikTok’s US operations moved into a new majority-American joint venture in January 2026, which removed the immediate ban risk but not the structural risk, and platform-specific rules on political content, health claims and AI labelling now differ enough that a single approval process will get something wrong. Our social media marketing guide covers the platform-level detail, and the video and creator economy news for September 2026 tracks what changed this month.

Connected TV and streaming ads

Connected TV is the part of video that grew up fastest. Netflix told advertisers at its May 2026 upfront that its ad tier reached more than 250 million monthly active users, up from 94 million in 2025, and said it would add the ad plan in 15 more countries including Ireland, the Netherlands, Poland, Indonesia and the Philippines. Amazon made ads the default on Prime Video. Disney, Paramount and Peacock all sell programmatically.

The catch is concentration. eMarketer’s 21 August 2026 forecast puts US CTV ad spend at US$36.95 billion in 2026 against US$3.45 billion in the UK and US$3.40 billion in China. If you are buying CTV outside North America, expect thinner inventory, higher effective CPMs and measurement that is a year or two behind.

Buyers are also unconvinced about quality in the open market. The IAB’s Part Two report from 14 July 2026 found 67% of video buyers express somewhat to no confidence in open exchange inventory, 55% for private marketplaces, and 43% even for direct and programmatic guaranteed deals. That gap is the strongest argument for buying CTV through publisher-direct or programmatic guaranteed deals rather than chasing cheap open-auction impressions. The mechanics of those deal types are covered in the programmatic advertising guide.

Live video and live shopping

Live is genuinely valuable and genuinely oversold. The IAB found 93% of video buyers agree live content commands a premium, 45% of them citing higher attention. But live shopping in Western markets remains a niche. eMarketer reported US livestream ecommerce sales of US$14.64 billion in 2025, up nearly 50%, but only 21.7% of US digital buyers had bought via a livestream and 43% of US adults said they had no interest at all.

So the sensible position is neither “live shopping is the future” nor “it does not work here”. It works where the category rewards it: collectables, beauty, fashion resale and anything where scarcity or personality drives the purchase. Whatnot and TikTok Shop prove that. A B2B software company running a monthly live demo is doing something different and should measure it as a webinar, not as commerce. The ecommerce marketing guide goes deeper on the shopping side.

Video SEO and getting found

Video search behaves in two ways at once. Inside YouTube the query is the ranking signal, so title, the first lines of description, spoken words in the first thirty seconds and the thumbnail carry most of the weight. In Google, video is one surface on a broader result page, and the winners are pages with useful text where the video is one element, not a bare embed.

Three things pay off reliably. Put a real transcript on the page, not just auto-captions in the player, because the text is what gets indexed. Use VideoObject structured data with a thumbnail, upload date and duration. And name the video after the question people actually type, which is usually longer and more awkward than the title a brand team would pick.

The bigger 2026 shift is that AI answer engines summarise, then link. A video with a clean transcript and a clear on-page summary can be cited by an AI system that will never play the video. That is a reason to publish the written version alongside every video you make, which is exactly the argument made in the SEO guide and the content marketing guide.

Production workflow and AI video tools

Generative video crossed from novelty into production this year. The IAB found nearly two thirds of digital video buyers using generative AI in creative work, up from about half in 2025, and expects roughly one third of ad assets to involve it in 2026, rising towards 43% by 2027. Wistia’s 2026 State of Video report, published 22 April 2026 from more than 13 million videos and 79 million viewing hours, found more than a third of teams already using AI in the workflow, concentrated in pre-production: scripting, planning and ideation.

That is the honest picture. AI is very good at the parts nobody enjoys (transcripts, rough cuts, versioning one ad into 40 sizes and languages) and still unreliable at the parts that carry your brand. Fully synthetic hero films remain rare for good reasons: talent rights, provenance labelling, and the fact that audiences notice.

On compliance, keep two habits. Label AI-generated or materially altered footage where the platform asks for it, because YouTube, Meta and TikTok all now require it and enforcement is automated. And keep provenance records for anything featuring a person, including consent for voice and likeness. The wider tooling picture sits in the AI in digital marketing guide.

Formats, hooks and length

Length follows the job, not a trend. Wistia’s data shows shorter videos hold a higher share of their audience, while videos of 30 to 60 minutes drive the highest click-through on in-player elements, because whoever stays that long is qualified. A 45-second product clip and a 40-minute webinar are both correct, at different stages.

Hooks matter more than production value. In feed video, the decision to keep watching is made before the logo appears, so lead with the outcome or the problem and put the brand in the middle. For in-stream ads, assume the skip button and design the first five seconds to work as a complete message on its own. For CTV, assume the opposite: nobody skips, but they may be looking at a phone, so carry the message in the audio.

Three formats are underused. Customer objection videos, which take the exact sentence prospects say on sales calls and answer it in 60 seconds. Comparison videos, including honest ones that name competitors. And process videos showing how the thing is made or delivered, which build more trust than any testimonial because they are hard to fake.

Benchmarks and costs

Benchmarks are sanity checks, not targets. Video costs vary more by country, category and season than by anything you control. Use the table below as a starting point, then replace every row with your own trailing 90-day data.

Treat published CPM and cost-per-view averages with suspicion unless the source names its dataset and date. Most quoted YouTube figures circulating in 2026 come from agency samples, not from Google. For a defensible planning number, run a two-week test at low budget and use your own auction data. The bidding mechanics sit in the PPC and paid search guide.

Measurement that holds up

Video measurement has one persistent problem: the platforms that sell the impression also grade the homework. View-through conversions inflate, last-click undercounts, and CTV rarely produces a click at all. The fix is not a better attribution model. It is a combination of methods.

Use platform reporting to optimise inside a campaign, because it is the only signal the algorithm responds to. Use geo holdouts or lift studies for the questions that decide budget: a two-week blackout in matched regions tells you more about CTV than any dashboard. And use mix modelling once spend supports it, which means seven figures a year across several channels. The marketing analytics guide covers the modelling side.

One more discipline. Agree the counting rule before launch, in writing. Half the arguments about whether video worked are arguments about whether a two-second autoplay counts as a view.

A 90-day video plan

For a team starting from zero, with one editor and a modest budget, this order works.

Weeks 1 to 3. Pick ten questions your sales or support team answers every week. Record ten videos answering them, one take each, no set. Publish them on your own site with transcripts and VideoObject markup, and on YouTube as long-form.

Weeks 4 to 6. Cut those ten into 30 vertical clips. Publish across Shorts, Reels and TikTok, three a week per platform. Measure hook rate only. You are looking for which openings survive, not which topics.

Weeks 7 to 9. Take the three best-performing clips and put paid budget behind them on one platform. Add a demo or testimonial video to your highest-traffic conversion page and measure the change in conversion rate, which is the fastest financial return in this whole plan and the reason the CRO guide is worth reading alongside this one.

Weeks 10 to 13. Run one creator collaboration and one small CTV or streaming test, each with a defined holdout. Review everything against cost per incremental conversion, then rebuild the next quarter’s plan around whichever of the two produced a measurable lift. Creator pricing and contracts are covered in the influencer and creator marketing guide, and the broader channel mix in the complete digital marketing guide.

Last researched and updated: 7 September 2026.