Are HTML5 Banners Still Worth It in 2026?

"Banner ads are dead" has been a headline every year since 2014. Meanwhile, every media agency I work with still sends me a brief every week with the same six sizes in it. Somebody is buying those impressions. Here is why HTML5 banners are still worth building in 2026, when they are not, and what changed.

What actually pays for banners today: targeting, not the banner

The 2010s banner was a billboard: the same 300×250 for everyone, bought by the thousand. That banner deserved its bad reputation.

The 2026 banner is bought programmatically against an audience: people who abandoned a cart, people in a postcode, people who watched 75% of a video last week, people whose weather is currently rain. The creative is the last mile of a targeting decision that has already been made. When targeting is good, a modest banner performs; when targeting is bad, no creative saves it.

This is why the format survived: display is the cheapest way to put a specific message in front of a specific person, at scale, on the open web. Social has walled gardens; search has intent but no reach; video has attention but costs ten times more per impression. Display is the glue in between.

Why HTML5 specifically, and not just a JPG

A static image is fine when the message is one thing for everyone. HTML5 earns its place the moment you need any of these:

1. Dynamic creative. One template, thousands of variations rendered at serve time: the product the user looked at, the price in their currency, the store nearest to them, the countdown to the end of the sale. You cannot do this with a JPG. Dynamic creative optimisation (DCO) is the single biggest reason HTML5 is still the default in retail, travel and automotive.

2. Interactivity. A car configurator in a 300×600. A quiz. A carousel the user swipes. A "find your size" mini-form. Interaction rates are small in absolute terms, but an interacted impression is worth many passive ones, and it is measurable.

3. Motion that respects the weight limit. A 15-second animated story in a 150 KB zip, crisp on retina screens, versus a 2 MB GIF that looks like a 2009 forum signature. HTML5 with vector shapes and code-driven animation is the only way to get real motion under the limits publishers enforce.

4. Localisation at scale. Twelve languages × six sizes is 72 static exports and 72 files to re-export every time legal changes one word. In HTML5 it is one build with a text file per language.

5. Live data. Weather, sports scores, stock, delivery windows. Any banner that says something true right now has to be code.

When a static banner is the right call

Be honest with the client about this, because they will thank you when the invoice arrives:

• The message is one sentence and one image, and it will not change during the flight.
• The placement is a strict publisher with a 40 KB static-only spec (still common in native-ish and email placements).
• The budget is small and the campaign is short; the production cost of HTML5 would eat the media budget.
• It is a test. Ship static, learn what message works, then build HTML5 around the winner.

A good static banner beats a lazy HTML5 banner every day. Animation is not a strategy.

What changed in the last few years

• Cookies went away, targeting did not. Contextual, first-party and cohort-based targeting took over from third-party cookies. The creative side is unchanged: you still deliver a zip.
• Google Ads made HTML5 harder to get into (90 days, $9,000 spend, clean history) and easier once you are in (600 KB limit instead of 150 KB). See the 2026 HTML5 specs per platform.
• Sizmek/Amazon Ad Server is gone, so the EB.clickthrough generation of banners is now legacy code. Most production converged on the CM360-style clickTag and Adform's manifest.json.
• AI-generated statics got very good. Which, counter-intuitively, made HTML5 more valuable: when everybody's static looks polished, motion, data and interaction are what remain hard to fake.
• Attention metrics (viewable time, in-view animation completion) are replacing CTR in reporting. Animated HTML5 banners hold attention longer than statics, and now that shows up in the dashboard.

The unglamorous truth about banner production

The creative is rarely the bottleneck. The bottleneck is the twelve emails after the creative: "which size is this", "the client can't open the zip", "is this the latest version", "did legal approve the German one". Most banner projects lose a day to logistics for every day of design.

That is a solvable problem, and it is exactly why I built BannerPeek: drop all the zips, images and videos of a campaign, get one link, the client sees every size running live, clicks on the banner to comment, and approves with a name and a timestamp. There is a free plan, no credit card.

Short answer

Yes. HTML5 banners are worth it in 2026 whenever the message needs to change per person, per place, per moment, or needs motion under a real weight limit. Statics are worth it when it doesn't. The skill is knowing which brief is which, and not letting the delivery process cost more than the creative.