Co-design: from expert- to user-driven ideas in public service design

Figures & data

Table 1. Overview of the cases, the co-design process, and the participants.

Project & Aims	Resourcing & Planning	Recruitment	Sensitization	Facilitation
Project A: Alcohol education program development for parents of high-school students in Queensland, Australia.	A systematic literature review (SLR) was undertaken to understand best practices concerning alcohol education for parents. A total of 13 program activities were sourced and featured as activity cards during the co-design sessions.	Parents were recruited via a network contact’s restaurant Facebook page. A total of 24 participants (15 female; aged 38–68) who had at least one child aged between 12 and 16 were recruited.	A brief introduction was provided to the underlying topic and the design task at hand. Researchers assigned people to groups, but then gave them time to individually reflect on each of the activities and to categorize these into likes and dislikes. Next, a moderated group discussion allowed participants to share their thoughts with others.	Two co-design sessions (90–120 minutes) were conducted as part of the ideation stage of program development. Participants collaborated in teams of 4–5 and 3 facilitators ensured that teams stayed on-task. Groups received butcher paper, pens, coloured markers, and post-it notes. The teams pitched their developed programmes to the research team and to other participants.
Project B: Virtual reality (VR) alcohol education content development for high-school students in Queensland, Australia.	An SLR was undertaken to understand current VR use in alcohol programs. None of the identified programs were deemed suitable for inclusion in the co-design session. Different sensitizing stimuli were planned for the co-design sessions.	A convenience sampling technique was used from existing networks. A total of 21 Australian high-school students (9 female; aged 14–16) were recruited. Sixteen indicated that they had already tried/consumed alcohol previously.	Students were asked to fill in a brief survey containing questions on general demographic data, and their perceived peer pressure susceptibility. Following a brief introduction, each group received different sensitizing materials; 1) role play, 2) video exposure, and 3) an online simulation.	One co-design session (60 minutes) was conducted as part of virtual reality script development. Participants collaborated in teams of 5–6. Groups received butcher paper, pens, stickers, coloured markers, and post-it notes. The research team members ensured that the groups remained on-task.
Project C: Social media campaign content development for addressing alcohol-related violence in Brisbane, Australia.	Grey literature was searched through to identify programs that had been implemented previously in Australia to reduce alcohol-related violence. A total of 15 activity cards were created and used during the co-design sessions.	A total of 30 participants were recruited online via the university’s media channels and the commissioning body’s Facebook and Twitter channels. Participants were also recruited through university lectures, the research team’s personal networks and a local bar.	Participants were provided with activity cards summarising existing campaigns and were then asked to individually sort them into piles containing likes and dislikes. A group discussion followed to gain insights into campaign effectiveness, target audience, improvements and what one campaign did well that another did not. Lastly, the participants selected the most- and least-liked campaign.	A total of four co-design sessions were conducted (90 minutes) for the ideation stage of a social media campaign addressing the prevalence of alcohol-related violence. Participants were split into teams of 3–4 and provided with pens, coloured markers, post it notes and butcher paper. Participants were encouraged to express their ideas using words and/or drawings.
Project D: Development of a campaign to reduce household food waste under Redland City Council (RCC), Australia.	An SLR and grey literature search was undertaken to identify previous food waste programmes, including those implemented in Australia. Nine food waste programmes were identified and incorporated into co-design planning.	A convenience sampling technique was used to recruit participants via an invitation email. A total of 21 participants were recruited from a large Australian university and among RCC employees-	Three sensitising activities were undertaken to prepare the participants: 1) 11 words were tested in a word association test. 2) A group discussion guided by 6 semi-structured questions. 3) A discussion about past campaigns to identify what participants liked (and did not like).	A total of two co-design sessions were conducted (45–60 minutes) to develop new ideas for a food waste reduction campaign. Participants were grouped into teams of 3. Teams received pens, coloured markers, post it notes and butcher paper. The teams were asked to pitch their final ideas to the research team.
Project E: Co-design of a weight management program in Queensland, Australia.	Grey literature was searched through to identify programs that had been implemented previously in Australia regarding weight management.	Project partner NGOs recruited overweight and obese respondents in local shopping centres offering them a free health check followed by an invite to a co-design session. A total of 26 participants attended the co-design sessions.	Three sensitising activities were undertaken to prepare the participants: 1) 10 words were tested in a word association test. 2) A 10-sentence reflection task was completed by each participant. 3) A discussion about past campaigns to identify what participants liked (and did not like).	Eight co-design sessions (120 minutes) were conducted. Small groups (1–5) were challenged to develop an idea or a campaign they felt would engage and motivate them as regards losing weight. To spark their creativity, various materials were provided (pens, pictures, glue, scissors, and large blank sheets of paper). Teams were asked to explain their final ideas to the research team.
Project F: Development of a train-the-trainer program to reduce dog and koala interactions under Redland City Council, Queensland Australia.	An SLR was undertaken to identify interventions aimed at reducing wildlife and domestic pet interactions. Expert interviews (n = 14) were additionally used to gain further insights.	A survey was distributed via a council’s media channels (email, social media, and newsletter) and a university’s email list. A total of 635 dog owners responded. Invites to participate in co-designing were sent to participants reporting dog-koala interactions. A total of 41 dog owners attended the co-design sessions.	Participants individually reviewed past campaigns to encourage and reflect on how dog-koala interactions could be reduced. Small groups were formed to further discuss the effectiveness of existing campaigns.	Three co-design sessions (90 minutes) were conducted with dog owners. Teams of 3–4 were formed and challenged to design a programme they felt would engage and motivate them as regards minimising dog-koala interactions. Using materials provided (pens, pictures, glue, scissors, and large blank sheets of paper), teams developed a campaign and were asked to present their final ideas to the research team.

Table 2. Overview of exemplary data and data analysis procedure.

Raw data example (Project B)			Raw data example (Project C)			Category code	Theme	Framework development
A literature/web search was conducted to identify programs that had been implemented to reduce alcohol-related violence. Eleven programs were identified of which all were expert-driven in their design. An example of a social media campaign aimed at reducing the prevalence of alcohol related violence:			A systematic literature/web search was conducted to understand how virtual reality (VR) is used in alcohol education. Online videos and professionally-developed roleplays, enacting the consequences of alcohol consumption, were used as inspiration when developing a video script for a VR alcohol education program. See below for a screenshot of an interactive video aimed at reducing adolescent drinking:			Expert-driven Lack of user-insights Far from reality Not in context Lack of technology One-sided perspective	Problem definition may be too narrow Resourcing seems to be expert-biased New design possibilities may evolve from technology	Resourcing: Need to understand the issue from a multi-actor perspective Problem space should be kept open to allow for alternatives
Card-sorting activities were prepared for sensitization by using the material from the resourcing stage. See below for examples of two cards (from a total of 15) which were used during the sensitization stage:			Videos and roleplay scripts were prepared for high-school students to both enact in class and reflect on. See below for an excerpt from a roleplay script used on female students:			Many sophisticated design methods Little direction from resourcing Difficult to plan ahead High level of uncertainty Multiple changes to planning Scenario planning	Methods and materials need to be short and adaptable Flexibility is key to planning Every project has its own challenges Unforeseen challenges require iterations	Planning: Background to the topic and the participants involved can affect the planning process Flexibility and agility help when preparing for the unknown
Recruitment using a number of channels to reach potential participants (social media channels of a university and the funding organization, direct emails to students, a media release). Very low response rates led to co-design sessions with few participants:			Recruitment by collaborating with a school to reach suitable students. The school took on the responsibility for recruiting the students. This, however, also meant that the school set the rules, leading to the challenge of facilitating one co-design session with a large number of participants within a 60-minute timeframe:
Session	Date	Total Participants	Session	Date	Total Participants
1	June 2016	6	1	September 2016	22
2	August 2016	4
Participant numbers improved substantially after recruitment was conducted collaboratively with a local restaurant’s social media site:
Session	Date	Total Participants
3	September 2016	10
4	September 2016	10
Participant comment during the group discussion following the card-sorting activity: I do kind of think about I’m angry and I’m just going to knock him out. Like, no way, like, that’s stupid, like, punches aren’t bad, you know, it’s not – it’s not like what you see in, like, a lot of, I guess, movies or media or whatever and it’s just, like, people just knocking each other about. It’s so much – it could be so much more dangerous than that and, like, I hadn’t even really thought that one punch can kill. So, it did make me think. (Male, 21)			Excerpt from a facilitator’s field notes commenting on the sensitizing activity whereby participants were asked to enact a roleplay script and to reflect on drinking scenes during which group pressure led to the overconsumption of alcohol: The roleplay got everyone excited and straight into acting out the peer pressure drinking scene (they were doing this standing up and engaging at a high level – it was difficult for me to stop them and remind them to move to the next step). Transition from roleplay to idea generation and critical reflection on peer pressure scene worked extremely well in this group because they really seemed to get the message. They started to exchange first ideas on how the roleplay script could be improved to make it more engaging for teenagers. (Group 1–4 males – Observation of roleplay activity using beer goggles which simulate drunkenness)			Raising awareness Users take ownership Building trust and confidence Engagement in topic Short but powerful Engaging activities Prepares non-designers	Sensitization triggers reflection Participants become engaged and start to identify with the topic Sensitization seems to help break down barriers Sensitization activities should be interactive and short	Sensitization: Activities should be geared towards triggering reflection and building confidence among participants Sensitization is the first step towards removing the expert from design
Excerpt from a facilitator’s field notes during the co-design session (this group showed high levels of engagement during the card-sorting activity): Group used one existing solution as starting point and developed the idea further by adding additional components. The group found consensus regarding what should be developed early in the process (within the first 5 minutes). The group members became more creative during the process and all members contributed their experiences/skills/knowledge. Little facilitation was required by the facilitators. Group cohesion and collaboration were strong – no disengagement by individuals or intra-team conflicts were evident. It seems that group members allocated specific roles to each other to work more effectively. Everyone seemed to know what to do.			Excerpt from a facilitator’s field notes during the co-design session (this group showed high levels of engagement during the card-sorting activity): Students turned to the script provided and started discussing what they liked and disliked about it. They could not identify any positive aspects despite repeatedly being asked by the facilitator. One participant’s reflection was that ‘the peer pressure’ aspect does not come across well. Within the group, one male participant took on a leadership role in discussing and developing the initial ideas on paper. He started drawing a garden party scene. The other participants provided inputs, including ideas for possible peer-pressure scenarios. Intra-team collaboration was facilitated through the discussion/leadership role taken by the one male participant. It appears that team members took on different roles during the process including questioning, scribing, providing feedback, and analyzing. (Group 2–2 females, 2 males – Observation of intra-team processes)			Experience sharing Participant empowerment Team-building challenges Collective creativity Experts as facilitators Existing solutions as inspiration Activating disengaged users Flexible facilitation approach	Experience sharing breaks barriers Enabling and supporting the novice teams in forming and ideating Aligning facilitation with the needs of the team	Facilitation: Facilitator carries the responsibility of enabling teams to maximize their creative collectivity Facilitating co-design teams means adjusting the facilitation approach to the needs of the diverse teams and their unique dynamics
(Group 2–3 females, 1 male – Observation of intra-team processes):			Additional notes: This project utilized three different sensitizing activities before the co-design activity (watching and reviewing a video, playing an interactive online game, and enacting a roleplay script). The teams more actively engaging in the sensitizing activity required less guidance during facilitation (e.g. Group 1), while the teams that were more passive during sensitization required a lot of encouragement from the facilitators (e.g. Group 3). This finding supports the relevance of applying engaging sensitization activities. The school did not give permission for pictures to be taken during the co-design activity.
Participants describe how their developed idea differs from the expert-driven ones, which were discussed during the sensitization stage: I like the idea of taking a more positive approach and saying, ‘Be logical. Be smart.’ It doesn’t have the negative – this doesn’t have the negative connotation behind most of the other messages. Saying to young adults ‘don’t drink at all’. That’s not going to work. So, I think, yeah, telling them to just, you know, make wise choices and to think about what they do is a better approach at times.” Idea ‘Don’t be a porkchop’A humorous social media campaign which depicts a situation during a night out when Shazza and Bazza are drinking heavily while Dazza is drinking responsibly. Bazza starts acting aggressively when someone accidently bumps into Shazza. Dazza intervenes and says ‘Don’t be a porkchop Bazza, don’t be that guy mate’. Dazza gets Bazza and Shazza some water. Shazza, Bazza and Dazza move to the dance floor and continue having a fun night.			Participants describe how their developed idea differs from the expert-driven ones, which were discussed during the sensitization stage: The content and the storyline were pretty good but the only fail, it was way too, like [pauses], formal, that’s the word it was like someone in their good twenties, thirties wrote the script and they didn’t really have a connection. It wasn’t like they had a connection with our generation so it made it [the script] very bland … like a lot of things, like, I bet ya half the class doesn’t even know what introvert means. Idea ‘House party’A comprehensive guideline regarding how a ‘real’ house party looks. Includes a number of recommendations for a peer pressure scene that is typical of a house party, including atmosphere features, specific background actions, a heterogeneous crowd in terms of age and gender, and the language used by the actors.			Relatable experiences Aligned with the user’s life Close to context Little concern regarding feasibility Improvements to existing solutions New perspectives and ideas regarding problems Too different means rejection Reflection on the original problem	Users contribute context-related insights Ideas generated seem to be more aligned to the users’ needs Users can have very different views of the same problem Ideas are not market-ready solutions Stakeholder involvement may increase support for implementation	Reflection and building for change: Moving user-driven ideas forward requires reflection and open dialogue with those responsible for implementation Implementing user-driven ideas means rethinking how service design was done previously

Figure 1. Example of cards used in Projects A, C and D.

Figure 2. A co-design framework for public service design.