Beautiful imagery has its place in getting your designs – and you – noticed. But do mockups full of models take a design where it needs to be? If the design is just a concept to pique attention, perhaps. If it’s meant to be a workable interface for real users, this can trap you into thinking too narrowly.

Choosing the right photos, especially those photos that are meant represent your users themselves, can be powerful tools in a user-centered approach to design. Don’t miss this opportunity to explore the questions they raise. They’ll ground your designs in (beautiful) reality.

Who are they are really?

To use realistic images, you need to find our who is really using your product. Ask the people on your team who may know: product managers, marketing and sales colleagues, data scientists, and especially previous designers.

If you’re working on a product that already has a set of personas, those may already have photos you can use. Using existing personas can save you some effort in tracking down new images. They are also a useful shorthand to a team that is already familiar with them.

You may also have representative users that have already been profiled for previous product work, including sales case studies, focus groups, marketing materials, or as participants of an alpha team.

While personas and representative users are good for understanding motivations or roles, they are just one data point. Depending on the extent of your member data, you may be able to get a more complete picture of your true user demographics: the distribution of ages, gender, and location of your users. You can use this kind of information to challenge yourself to include more representative samples of people in your designs.

How common are photos?

If you’re designing for an existing product, you should be able to find out some details about what happens with photos today. How many people use a profile photo? Do they add one when they create their account, or later on? How frequently do they change it? Find out how many people choose not to add an photo, and track that metric so you can determine whether your design changes are making it easier or more compelling to do.

Reminding your users to add a profile picture is a good idea, but you may get more leverage out of ways to help users begin with one from the start. Offering an option to sign up with an existing account is a good low-friction way to start with an image since they’ve already gone through the effort to pick a good picture. If your product is a mobile app, make it easy to take a new photo or choose one from their photo library during their setup process.

What will users see otherwise?

For profile photos, the no-image case is a good opportunity to think about (and potentially redesign) your default image. Whatever you choose for defaults will be prevalent in your product, and should be accounted for in your designs.

If you’re redesigning your default, consider how you can reinforce the brand or mood of your product. A single silhouette can easily skew too masculine, or end up looking alien or toylike in its attempt to be ambiguous or neutral about gender. Gendered defaults can help with that, but take care using those unless users can explicitly choose how they wish to be represented.

A logo or object can be a reasonably neutral choice. It can be a way to express a brand either directly or indirectly, such as Twitter’s anticipatory egg. It can also express personality, such as this thoughtful exploration from Jamie Dihiansan for Basecamp. Both of these take advantage of varying colors or defaults to provide a little more variety and interest.

The goal should always be to encourage true expression, and we set the tone about who should use our products with the people we show in them. Seeing friendly faces with a representative range of age, race, and gender, and even more personable defaults, answer the questions about your product’s intent. They will lead to a product that is more welcoming, and more real.