BACK TO ALL ARTICLES
July 24, 2026
Table of Contents
Wrong question.
That sounds glib, but stay with it: the mobile-versus-desktop framing is why a lot of display budgets are allocated poorly. Neither device wins. The sizes that carry most campaigns are the ones that work in both places, and the 300×250 is the obvious example, which we’ll get to. What genuinely distinguishes a phone from a desktop is not format performance. It’s what the person is doing when your ad loads. Someone thumbing through a news feed on a train has about a second of attention and no intention of filling in a form. Someone at a desk with three vendor tabs open is in an entirely different state of mind, and they will absolutely fill in the form, because that’s what they came to do.
Build around that, and the size decisions mostly make themselves. Here’s what earns its place in 2026, where each device actually pulls ahead, and the specs your creative has to clear before any of the rest of it matters.
Mobile carries the impression volume in 2026. That much is settled, and the formats built for it, the 320×50 and the 320×100, stopped being optional some time ago.
Between the two, the 320×100 usually earns the extra height it requires. Double the vertical space buys you a headline that reads at a glance and, more importantly, a tap target someone can hit on purpose. Accidental taps are not free. They burn budget, they produce nothing, and they quietly inflate your click numbers so the campaign looks healthier than it is. The 320×50 still has a job, mostly as a sticky bar pinned to the top or bottom of a screen, where being small is the entire point.
What mobile does well is discovery. It reaches people in the gaps of their day. First touch, brand introduction, app installs, and building the remarketing pool. Ask it to also close a complex purchase, and you’ll be disappointed, which is where many campaigns quietly go wrong. Motion earns its keep in some mobile placements and wastes budget in others we break it down in animated GIF vs. static display ads.
Fewer impressions. Much better conditions.
A person at a desk has a keyboard, a big screen, and usually a reason to be there. They open comparison tabs. They read the pricing page properly instead of pinching to zoom. They fill out the fourteen-field form, which would be genuinely painful on a phone. For B2B, enterprise software, anything with a real consideration cycle, that’s where the conversion happens, and no amount of mobile optimization changes it.
The formats follow. The 728×90 leaderboard holds the above-the-fold slot and gets you seen before the first scroll. The 300×600 half-page is the most creative space you get in a standard unit, enough for an actual product showcase rather than just one line of copy and a button. The 160×600 skyscraper trades impact for endurance, sitting in the sidebar through a long scroll. And the 300×250 works here too. It keeps coming up.
Starting from nothing? This set does most of the work:
That covers the large majority of GDN and programmatic inventory. Going beyond it hits diminishing returns fast, and truthfully, six sizes built carefully will outperform fifteen built in a rush every time.
Responsive should be your base layer. You hand over images, headlines, descriptions, and Google assembles them to fit whatever slot exists, including plenty your fixed sizes would never fill. Feed it properly, several images, a real spread of headlines and descriptions, or it has nothing to test and the whole mechanism stalls.
Now the limitation, because there is one. You lose control of the layout. The algorithm decides what fits, and on a premium placement where the impression cost is high and the composition genuinely matters, that’s a real cost. Hand-built static still wins there, because you decide where the logo sits, where the headline breaks, how much white space the button gets. We dig into that control-versus-automation trade-off and when a hand-built ad is worth it in responsive vs. uploaded display ads.
So: both. Responsive for coverage, static for the placements worth designing on purpose. And if your brand standards are strict, switch off the automated asset enhancements before you find out the hard way that the system has recropped your hero image in a way your brand lead will not enjoy explaining to anyone.
One more thing on this. Refresh the creative every four to six weeks. Fatigue is real, it’s measurable, and the same image running for four months turns into wallpaper while your acquisition cost drifts upward, and nobody notices because the impressions keep serving.
An ad nobody sees is worth nothing. Doesn’t matter how well it’s designed.
The MRC and IAB standard for a viewable display impression is 50% of the ad’s pixels in view for one continuous second. Campaigns that don’t track against it are frequently paying for impressions that never had a chance. There’s a wrinkle worth knowing, too: ads over 242,000 pixels only need 30% in view for that second. Practically, that means a 970×250 billboard gets measured on a different threshold than the standard units sitting around it, which occasionally explains a viewability report that looks stranger than it should.
Placement quality is the other half of this, and it’s the less glamorous half. Low-quality sites and apps generate impressions and accidental taps that appear as activity in a dashboard, while delivering nothing downstream. Pulling placement reports and excluding the persistent offenders is tedious. It also works reliably, which is why it stays part of running a display account properly.
Size ought to follow the objective, not a template someone saved in 2021.
DTC and consumer ecommerce: weight toward mobile. The 300×250 plus the mobile banners, the budget skewed to reach. B2B and considered purchases: hold real desktop spend, focus on the 300×250, the 728×90, and the 300×600, because that’s where somebody actually completes the form.
Awareness campaigns want a bigger canvas, so the 970×250 billboard and the 728×90. Direct response usually does better with the in-content rectangles, which sit where the reader’s eye already is rather than demanding it move.
The framing that helps most is sequence, not split. Mobile introduces the brand and fills the remarketing pool. Desktop re-engages those people when they’re in a position to decide. Treating the two as competitors for the same dollar misreads how the journey runs, and it’s a common enough mistake that it’s worth saying plainly.
This part isn’t a design detail. Miss the spec and the asset gets rejected, full stop, and heavy files slow pages down, which hurts viewability and makes publishers like you less.
The numbers that matter:
Compress before you upload, not after a rejection email. For the full breakdown by format and placement, Linear keeps a Google Ads display ad sizes cheat sheet checked against current Google documentation. Hand it to your designer before they start exporting, not after.
Blended reporting hides things. A campaign that looks middling in aggregate is often a mobile segment doing well and a desktop segment dragging, or the exact reverse, and you can’t act on either until you pull them apart.
Look at CTR and conversion rate separately by device. Mobile routinely posts the higher CTR while desktop posts the higher conversion rate on anything considered. A strong mobile CTR alongside weak mobile conversions usually points to form friction, not creative issues, and swapping the banner won’t fix it. Check viewability by device too, since app inventory and desktop web behave very differently. And watch eCPM across both, because it tells you what you’re really paying for attention in each environment.
Signal failure and tighter privacy rules pushed contextual relevance back to the middle of the display conversation. That raises the value of creative that fits where it lands. When you can’t lean on granular behavioral targeting anymore, the ad itself has to carry more weight, and a good creative at the right size on a relevant page is the thing that keeps working.
Which is the argument for all of the above: a disciplined size set, specs that clear, performance read by device. None of it is clever. It’s just what separates campaigns that scale from campaigns that stall.
The 300×250, placed in-content because it sits where focus already is. For dedicated mobile banner slots, the 320×100 beats the 320×50 in most cases, mainly because the extra height gives you a readable headline and a tap target people can actually hit.
Depends what you’re optimizing for. Responsive wins on reach and on production time, since one asset set fills placements of no fixed size. Fixed wins on control, which matters more than people expect on premium inventory. Most accounts should run responsive as the foundation and add static where the placement justifies the design work.
The 300×250, the 728×90, and the 300×600 are weighted toward desktop. B2B buyers research and convert at a desk, and those three give you room to make a case that a complex purchase needs.
Four to six weeks. Repetition produces banner blindness, CTR slides, acquisition costs climb, and because impressions keep serving, the whole thing decays quietly rather than failing visibly.
Uploaded static image ads cap at 150 KB. Responsive display and Performance Max image assets go up to 5 MB. Accepted formats are JPG, PNG, GIF, and HTML5, and animation is capped at 30 seconds with a 5 fps ceiling on animated GIFs.
Build sizes that cover real inventory. Meet the specs so the creative actually serves. Run responsive for reach and static where control matters. Read results by device rather than in aggregate. That’s most of the job, and it’s less complicated than the amount written about it suggests.
If you’d rather not do it yourself, Linear’s PPC management services put dedicated specialists on your account. No account-manager layer in between.
Using data collected from our in-depth audit, we’ll deliver a detailed plan to grow your business month after month. Your proposal includes:
WRITTEN BY
Luke Heinecke
Like what you read?
Your free proposal is overflowing with improvements.
RELATED ARTICLES
Display Ad Types and Formats: A Practical 2026 Guide
Choosing a display ad is not simply a matter of selecting a banner size. You are making several conn...
August 14, 2026
How to Compress Display Ads Under Google’s 150KB Limit (2026)
Google rejects uploaded static display ads that exceed 150KB, so a banner that looks perfect in your...
August 5, 2026
Responsive vs Uploaded Display Ads: Specs, Control, and the Demand Gen Shift
A 400 KB banner that looks flawless in your design file will not upload to Google Ads. The same artw...
July 29, 2026
Made with 💙 in Utah