Most decks I get asked to fix don't have a design problem. They have a photo problem. Someone spent twenty minutes picking a "nice" image, dropped it on the slide, and never asked what job it was supposed to do. The photo looks fine in isolation. Next to the headline and the data, it fights everything around it.

Choosing photos for a presentation isn't really a taste exercise. It's closer to casting: every image has a role, and the wrong actor in the right costume still throws off the scene.

Start with the job, not the mood

Before opening a stock library, ask what the photo is doing on that specific slide. In my experience there are really only four jobs a presentation photo ever does:

  • Anchor — sets context for a section (an office, a factory floor, a city)
  • Evidence — shows something the audience needs to verify visually (a product, a defect, a before/after)
  • Breather — gives the eye a rest between dense data slides
  • Transition — signals a shift in topic or tone between sections

If you can't say which of these a photo is doing, it's decoration, and decoration is the first thing an audience's eye slides past — which defeats the point of spending time choosing it.

Match the crop to the content that sits next to it

One thing people overlook: a photo doesn't exist alone on a slide, it exists next to a headline, a chart, or a block of text, and the crop has to leave room for that neighbor without fighting it. A tightly cropped portrait facing the wrong direction — looking off the edge of the slide instead of into it — pulls attention away from your text every single time, even if the photo itself is high quality.

A rule I'd actually stand behind: if there's a face in the photo, it should look toward your text block, not away from it. This isn't superstition — it's basic eye-tracking behavior; viewers follow gaze direction, and you want that direction pointing at your point, not off the page.

Resolution math that people get wrong

Everyone knows "use high resolution," but the number that matters is native pixel width versus your screen's projected width, not the file size on disk. A 1200px-wide image looks sharp on a laptop screen shared over video call. The same image projected on a 4K conference room display, full-bleed, will visibly soften — not because the file is "low quality" but because you're asking it to stretch past its native resolution.

If you don't know the venue, size for worst case: full-bleed background photos should start above 1900px wide. Anything smaller and used as a small supporting image in the corner of a slide is usually fine at lower resolution — the failure mode is almost always full-bleed stretching, not thumbnail use.

Watch for photos that quietly contradict your data

This is the mistake I see most often in business decks specifically: a slide claims growth, efficiency, or scale, and the photo next to it shows the opposite — an empty office when the copy says "thriving team," an outdated-looking product shot next to a claim about innovation. Audiences read images faster than text. If the photo and the headline send different signals, the photo usually wins, and it undercuts the claim before anyone reads a single word of the body copy.

Before finalizing a slide, look at the photo with the text covered. What does it claim on its own? Then compare that to what the headline claims. If they don't match, the photo needs to change — not the headline.

Consistency of light direction and color temperature

Mixing a warm, golden-hour photo with a cool, blue-toned office photo two slides later reads as inconsistent even to viewers who couldn't tell you why it feels off. It's not about using identical filters — that's usually unnecessary — but light direction and color temperature should stay roughly consistent within a single section of the deck, especially if photos appear back-to-back. A quick fix: batch-select the photos for one section and do a light temperature pass together, rather than picking each photo in isolation.

When a hypothetical illustrates this better than a rule

Say you're building a sustainability report deck. One common mistake is using a lush, editorial-style forest photo as a background for a slide reporting an actual emissions number. The tone mismatch — glossy nature photography next to a hard metric — makes the number feel like marketing rather than reporting. A plainer, more documentary-style photo (or no photo at all, just the number, large) usually serves the actual message better. This is a hypothetical scenario, not a specific case, but the pattern shows up across sustainability, financial, and safety reporting decks alike: the more serious the data, the less it wants a "beautiful" photo competing with it.

Don't be afraid to cut the photo entirely

I wouldn't recommend defaulting to "every slide needs an image." Some slides are stronger with just type and whitespace, especially data-heavy ones where a photo becomes visual noise competing with a chart for attention. If you're choosing between a mediocre photo and no photo, no photo almost always wins.

Next time you're building a deck, try going through slide by slide and naming, out loud, what job each photo is doing. If you can't name it in one sentence, that's usually the slide worth revisiting first.