Making Service Recipes better: user research findings and next steps

Sharing what we learnt from 160+ insights and prioritising 38 how might we statements. Explaining what we’re doing with it. Written for people who took part in the user research and anyone else familiar with the platform.

The Service Recipes homepage, with a set of combined notes from 6 testing sessions.

Since September I’ve been running Service Recipes, a platform showing how charities have done digital things. It shows these things in ways that others can copy and reuse for their own services and internal processes.

In November we did a lot of research into the platform. I’ve already written:

• Synthesising 7 different types of research
• The Good, the Bad and the Very Interesting User Research
• A record of what we did, why and how on the Service Recipes Notion site.

The research

We carried out 7 types of research.

1. Analytics — mainly to guide questions rather than insights
2. Tags and patterns research — understanding how tags and patterns are used on the site and in other places on the web. Our insight statements from this will guide user testing on tags and patterns.
3. Comparison — comparing the product with others will generate ideas that we can introduce later on in the process
4. Heuristic evaluation — we’ve commissioned two of these — one UX and one content design. The insights will inform our testing focus areas.
5. Recipe content crits — we’ve reviewed 10 recipes using content design best practice. Some insights will inform testing while others will go into the big synth pot
6. User interviews — understanding 6 people’s experience of Service Recipes and sentiment towards it
7. Usability testing — 50 minutes each with 6 people testing search functionality, interface, tags and recipe content. Half this time was spent on 2–3 recipe reads per participant.

6 recipes, annotated with notes from content reads with 3 users. Users read the recipe and commentated on their experience.

The insights

We drowned in insights! We had so many from across the 7 research types that we had to run a ‘synth of synth’ session to group, analyse and condense them. Then we turned them into ‘How Might We…?’ (HMW) statements (see below) but for now here are 10 of the best insights.

1. People value service recipes. They have returned to use them a few times and are fans of the concept. However too often they don’t find the recipes they spend time on relevant or useful to what they were looking for help on.
2. People gain different types of value from service recipes. Awareness, inspiration, reassurance (that they have done something similar well enough), and sometimes they partially use ideas from multiple recipes or copy a whole one to help them with a similar service or task in their organisation.
3. People often share a recipe or part of it with others in their organisation. They do this to help them think about a digital challenge or show them what might be possible.
4. The search feature doesn’t offer a good experience. Results often don’t seem relevant to the search term. However people often find other useful or interesting recipes in the results returned.
5. People don’t notice the recipe tags. Instead they only use the recipe titles to gauge whether they want to read the recipe.
6. People like the ‘service patterns’ tags because they are verb based. However they find them difficult to use as a way to find recipes and information they want.
7. Recipe titles are often slightly inaccurate or don’t give people the information they need to decide if it is relevant to them
8. Recipe overviews are too long, focus too much on the organisation behind the recipe and don’t help people decide whether the recipe is going to give them what they are looking for. Sometimes they offer false hope.
9. Recipes in general are not skimmable enough. It is difficult to extract information when skim reading because readability is poor and the content pattern (the sequence of sections of content) doesn’t support it enough.
10. The information found in the ‘Steps’, ‘Guidance’, ‘Risks’ and ‘Tool considerations’ sections is often mixed up. For example sometimes guidance can be found in both ‘Steps’ and ‘Guidance’ sections. Risks can be found in ‘Risks’, ‘Steps’ and ‘Tool considerations’. This makes the reader have to work harder to find the information they want.

There was loads more, but for now, I hope that gives you a sense of what we need to deal with on the platform. In fact, the ‘How Might We…?’ statements cover these 10 insights and others. More importantly though we’ve prioritised the statements for working on in January and February 2023.

160+ insight statements, clustered during our ‘synth-of-synth’ or ‘meta-synth’ session that pulled together insights from 7 pieces of research.

The ‘How Might We’s’, prioritised…

I think the statements speak for themselves.

1. How might we help first-time users understand what recipes are so they feel confident to proceed?

Includes:

• How might we describe Service Recipes in ways that people relate to better?
• How might we better articulate the value that people get from recipes?
• How might we make sure that the home page provides enough orientation early on the page for a new visitor.

This will be purely a content task.

2. How might we improve the content pattern to make it more useful to users?

Includes:

• How might we resolve the issues around what appears in steps, guidance, risks and considerations?
• How might we connect tools better with risk/safety information?
• How might we highlight key information that most users are looking for?

This will need some content design and technical development to implement.

3. How might we better display metadata in a recipe?

Metadata is the key information related to a recipe — from date it was written to content tags to tools needed to implement the recipe.

Includes:

• How might we present metadata in a way that helps the user decide if the recipe is relevant for them?
• How might we provide more meaningful tools price/budget information?
• How might we replace service pattern categories with ‘words that they understand’ that the users relate to more?

This will need some content design, interface design and technical development to implement.

4. How might we change ‘search results’ and the ‘browse’ page to help people take a further step towards a recipe that will be useful to them?

Includes:

• How might we change the focus from ‘tags’ that people don’t use to ‘elements’ (e.g. org, title keywords) they do use to help them select a recipe?
• How might we create more consistency in recipe titles and improve the match between title and content?
• How might we improve the search so that results are a better match to the search term?
• How might we make use of age categories so they show specific to the age patterns rather than general patterns?

This will need some content design, interface design and technical development to implement.

5. How might we make sure that content style is consistent across recipes so that users can get information from them more easily?

Includes:

• How might we ensure that steps contain sufficient detail that the user can replicate?
• How might we make steps useful as single steps, not just as part of a recipe?
• How might we ensure language used is relevant, friendly and reflective of sector values?
• How might we increase skimmability of the recipe?
• How might we make risks detailed enough to be useful?
• How might we improve the images used so that they inform the reader rather than decorate the recipe?

This will need us to edit existing recipes, and some content design and technical development for new ones.

6. How might we value the fact that users are often interested to browse recipes that interest them for reasons unrelated to their original task?

Includes:

• How might we allow people to find more recipes that are related in some way to the one that has inspired them?

This will likely need some design and technical development.

7. How might we make the UI more user friendly?

Includes a series of small areas for improvement:

• How might we make sure that onward pathways from filters are clear if we continue to have a filter set?
• How might add more definition into the page to support dyslexia users in following grid layouts?
• How might we make sure any side menus are seen and clearly usable?
• How might we make sure that titles that are hyperlinks are very clearly titles?
• How might we make it clearer to the user when they have reached the end of the recipe?
• How might we make sure that people understand that the ‘helpful or not’ feedback buttons and the get help pathway are two different things?

These will all need some design and technical development.

8. How might we support users who are interested in the contributing organisation as a means of identifying and trusting recipes?

Includes:

• How might we help users find recipes from organisations they consider to be like them (because the ‘life events’ tag is not effective)?
• How might we develop the contributors page so that it behaves more as users expect it to (eg organisation links take users to a list of their recipes)?

This will need some design and technical development.

9. How might we improve the experience for people who use the get help form?

This will need some design and technical development.

10. How might we optimise SEO?

This will be a content design challenge.

What next?

We’ve started scoping the first 7 HMWs for an 8 day technical sprint beginning Jan 25th. Realistically we:

• won’t get everything we want to get done, done. That’s ok. Its why we’ve prioritised.
• will need to do a fair bit of editing of existing recipes to improve their readability and quality
• will need to test anything we’ve redesigned before it goes live (more user testing opportunities to come)

This means you might not see any live changes on the site until the end of Feb. By then we will be planning more dev for March or April. We should also be putting new recipes onto the site in a new format.

Thanks to participants

Thanks to the six who helped us out here, each spending 90 minutes with us. We paid them for their time and expect to test again, probably with a new crew, in Feb 2023.