If you ever attempted to build a website on your own, you would’ve realized pretty quickly that opening Unsplash and typing in the search phrase “happy office people” is not a universal strategy. Yes, sometimes that’s enough, but more often than not, it’s not believable to the consumer.
As is the case with most things in life, there’s more than one way to assemble convincing website imagery. But all of them usually involve these:
- Decide what the image has to prove before deciding where to get it.
- A product photo, hotel room, software screenshot and decorative blog header are doing fundamentally different jobs.
- The preparation stage varies too.
- A small business may need to resize photographs, change image formats, clean up supplier files.
- Use Convertio to convert png to ico when preparing visual assets for a website without installing additional conversion software. The same goes for any image extensions as Convertio supports hundreds of different formats and lets you move from one to the other seamlessly.
- You might deal with svg files at some point. If you find them outdated, there’s a way to avoid them, too. Use Convertio as an svg converter to get PNG, JPG, PDF or other formats as needed, and those files can be used in software or systems that do not support svg.
Nielsen Norman Group’s eye-tracking research supports the logic. Users tend to pay attention to images containing relevant information, particularly photographs of products and real people, while ignoring generic decorative imagery. In one example involving employee biographies, users spent considerable attention on photographs of the actual staff. Needless to say, generic stock-style images were quick to be dismissed. This draws the logical conclusion that the image source should follow the job.
An online shop and a law firm should not shop for images the same way
Let’s consider one of the most common fields, ecommerce.
Baymard Institute found in product-page usability testing that 56% of users’ first action is examining product images. Its research also found that 42% of users try to establish a product’s size from the pictures, which is why Baymard recommends providing at least one image showing the product in scale or in a familiar context. A technically excellent studio shot may therefore still be inadequate if the customer cannot tell whether the vase is 20 or 60 centimetres tall.
That same logic changes depending on the website.
| Website | Images that carry the most value | Sensible source |
| Ecommerce | Exact product, details, scale, variants, product in use | Original product photography, licensed supplier assets, customer/UGC images |
| Hotel / rental / real estate | Actual rooms, property, location, amenities | Original photography |
| Restaurant / food business | Actual food, room, staff, atmosphere | Original photography |
| Professional/local service | Real team, premises, work, before-and-after examples | DIY or professional photography |
| SaaS / software | Actual interface, workflows, features | Screenshots, screen captures, diagrams, custom graphics |
| Portfolio | Actual work | Creator’s own photography, artwork or screenshots |
| Blog / magazine | Illustration of a subject rather than proof of a product | Editorial photography, stock, public-domain material, commissioned illustration |
| Corporate/B2B | People, facilities, products, processes | Original photography plus diagrams and illustration |
| Concept-heavy brand/startup | Mood and visual identity | Commissioned illustration, 3D work, custom design or carefully chosen stock |
The search for the picture is only half the task
The file that the photographer sends you is not the file that should end up on the website. Not always.
The photographer can provide JPEG originals or TIFFs that are very large. A designer may send an SVG logo. Some manufacturers may supply product images in PNG format. There is no “best website image format” as screenshots, icons, transparent graphics and photographs all behave differently.
Google’s web.dev guidelines recommend using SVG for geometric graphics like logos and icons, and photos or similar raster images should be used in JPEG, WebP, or AVIF. Use WebP and AVIF for efficient compression, and PNG for situations where lossless raster output or transparency is needed.
Convertio can convert all of the mentioned formats, while making the interswitching quick and easy. Rather than opening professional graphics software for the sole purpose of converting a wrong file type, a company can use Convertio’s image-conversion tools in the browser to convert the image into a desired format for the subsequent step in their web process. The converted file can then be resized, compressed or further processed by the CMS or the website builder or by the image CDN.
Many big sites are already doing this automatically. Responsive image systems can offer various image dimensions for various screens, and modern delivery platforms can encode resources in various ways according to the capabilities of the browser. That is, the master image and the final file delivered to any visitor are not required to be the same.
Pretty pictures = heavy website
The HTTP Archive’s 2025 Web Almanac found that images accounted for approximately 1,058 KB on the median desktop homepage and 911 KB on mobile—more bytes than any other resource category on those median homepages. It also found that images were the Largest Contentful Paint element on 85.3% of desktop pages and 76% of mobile pages.
Thus, a beautiful hero photo can be one of the very first things a visitor encounters, and one of the biggest things a browser has to download.
Optimisation, then, should be at the heart of the image strategy, right from the beginning. A picture taken with a phone that is several thousand pixels wide doesn’t have to be sent at the original size within a small content card. Responsive images can allow the browser to select a size that is a good match for the resource, and the main image, above the fold, can be treated differently than the images, such as photos, halfway down the page.
The image should also be comprehensible beyond the visual design. Google suggests using descriptive names for files wherever possible, as well as surrounding text and alt text that is helpful in describing the image content. The W3C also differentiates informative images, which must have a meaningful text alternative, from purely decorative images, which should usually have an empty alt=”” attribute so that screen readers do not read them.
But this doesn’t imply that all companies require a photographer, art director, image CDN and a weeklong asset audit. Quite the opposite. A small business can take photos of their products with a mobile camera, add real staff photos, add from stock photos (if available), deal with odd file formats via Convertio, and have a decent website platform do much of the work.
It does not need to be a blank space that has to be filled with any image.
If a visitor needs proof, show the real thing. If they require an explanation, make a diagram or screen shot. If they require some atmosphere then stock may be just fine. And if an image is doing nothing more than filling in the space the template allocated in some random way, then maybe the least expensive and easiest image source is no image source at all.
