Down the Ladder, Shape It Clearer
Developing The Prototype of Post-purchase Consumer Empowerment
Lots of findings, insights and ideas have been generated throughout the past 4 sprints. As we progressed into Sprint 5, we dove further down into the ideation phase to begin envisioning what a preliminary prototype should look like, such that it encourages consumer to exercise their rights, and delivers moments of delight even after the point of purchase.
Feedback from Client Stakeholders
There is greatest design opportunity for subscriber retention in the post-purchase, service request phase through the use of a smart agent.
After presenting our insights — including results from the guerrilla interviews, reframed problem statement, and the semi-structured interview on customer service experience, we gathered valuable feedback from our clients which has been the driving force for Sprint 5 as we expand upon our focus:
- How would our CR agent come into contact with consumers? Could the interaction already happen during the unboxing experience — moment right after receipt of products/services, leading all the way towards service request?
- Consider both sides of the journey: what does a “happy path” for a consumer look like? How might we as Consumer Reports, in those moments of delight, interact with consumers to deepen the relationship and establish trust?
Defining “Where It All Begins”
Upon ruminating on client feedback, we determined the most crucial decision to make at this stage is to “define” — narrowing down the scope by pinpointing specific domain(s) on which our prototype will focus.
Impact/Difficulty Matrix
Borrowing from LUMA Institute, we employed the “importance/difficulty matrix” method to quickly visualize our options; domains in the lower right quadrant are our top considerations as they generate high impact with low difficulty of implementation or risk. These include Cars, Home Appliances/Improvement, Baby Products, Cable Services, Airlines, and Hotels— aligning well with Consumer Reports’ value prepositions.
Among these industries, we did a second round of voting to further narrow down into three domains where we would storyboard to imagine the unboxing experience:
- Home (appliances, services)
- Cars (self-driving cars, IoT capabilities)
- Travel (experiential services)
Storyboarding The Unboxing Experience
How might the CR agent build trust with consumers in the post-purchase phase so that they, upon moments of delight, can place trust in the agent before permitting it to act on their behalf should the need for service request arise?
A mind map was developed to summarize all the ideas we have in different scenarios, surfacing several patterns in the consumer-agent interactions that are foundational to building trust — which led us to make a diagram of agent intervention next.
Diagram of Agent Intervention Across Post-purchase Journey
While discussing the various ideas that emerged from the storyboarding exercise, we extracted certain post-purchase phase tasks we believed our agent could implement. We divided these tasks into three categories: Inform, Guide and Act.
In the inform stage, we explored tasks like: informing consumers of their rights, company policies, and high-quality, curated reviews from other subscribers. In the guide stage, we explored tasks like: giving consumers guidance on forms to fill, what questions to ask businesses, how to construct an argument, where to find customer service information, who to talk to etc. And in the act stage, we explored tasks like: filling out forms and filing claims on behalf of the consumer, follow-up with businesses and solve post-purchase issues etc.
We also realized that not all consumers will go through this entire agent intervention journey. It’s a funnel that will be dependent on the individual, their preferences and their trust in CR and technology. We believe that consumer engagement will decrease as agent intervention increases through this journey — a large number of users will use the agent for inform tasks, a smaller number that trust the agent a little more will let the agent guide them, and an even smaller number will trust the agent enough to let it act on their behalf. Providing a service that allows consumers to engage with the agent at different levels of intervention will help us capture a larger share of the consumer base. This may present an opportunity for us to provide a fremium version of the agent, wherein the inform and guide services are provided for free while the act service is paid. This will help bring in consumers that are not familiar with CR, while also providing greater value to existing CR subscribers and loyalists.
Thus, creating this model to narrow down the agent’s tasks helped us gain clarity on what our agent will do and we got a more concrete and defined understanding of the agent that we will be designing for our project.
While discussing this model in the context of the 3 domains we narrowed, we realized that the specific tasks and modality of the agent will largely depend on the domain and specific problem we are focusing on. As a result, we decided to further narrow down to one domain to start with, so we can prototype using high content fidelity such that it effectively suspends participant disbelief and provides deeper, more valuable insights. From group discussions, we concluded the Home domain would be a good place to start. This is because it is a domain that is popular amongst current CR subscribers and it is a domain CR is well-known for, so it has a competitive advantage in this space. Additionally, while Cars and Home both are popular domains for CR, another research group is currently working on a project in the Cars domain. Thus, we see greatest opportunity in the Home domain and we decided to move forward in the project designing an agent in the context of Home appliances and services, like kitchen oven, Wifi services etc. Additionally, given the rise in IoT presence in our homes, we have an opportunity to inform and help consumers advocate for themselves in their interactions with more complex product service systems that introduce new risks and challenges.
Our goal now is to design an agent that assists consumers and empowers them to advocate for themselves in their post-purchase experiences with businesses offering home goods and services.
Developing “How It Works”
Having defined and reframed our problem space informed, now we finally move forward into the prototyping phase where we flesh out how the CR agent actually takes shape and form in the context of Home products/services purchase experiences.
Prototype Planning with Spider Map
To begin our developing phase, we mapped out our considerations in the 5 dimensions of the prototype to prioritize dimension(s) that warrant higher level of details.
By weighting different dimensions on a scale of 1 to 5, we determined that content and breadth are 2 of the most important aspects in our prototype.
- Content: The content in our prototype needs to be high fidelity because we need the agent to conduct specific tasks and provide scenarios that fit within the context of the post-purchase experience of home goods and services. Being as specific as possible with the content in terms of the tasks the agent conducts, the scenario, and the language used by the agent will help us effectively suspend participant disbelief, hence providing us with deeper, more valuable insights from testing.
- Breadth: A variety of tasks within the post-purchase phase of products/services in the Home domain should be covered to allow us to identify which ones are considered more useful/urgent before further refining the depth of each task.
- Depth: Once we have prioritized the tasks that the agent will conduct, we can then enhance the depth of certain interactions where higher demand/pain point is identified.
- Interaction: Higher fidelity in interaction will be realized further down the line by enhancing feedback & feedforward to deliver status and progress as part of the trust building paradigm.
- Visual: This is given the lowest importance as we will mostly be leveraging the existing design system from Consumer Reports to implement the interface of our system.
Personas & Jobs to Be Done
In order to better define the tasks that the agent will conduct, we recognized a need to better define the different personas that will interact with the agent and determine their specific Jobs To Be Done. As a result, this week, we spent time creating personas together. We created 3 personas — CR loyalist, Gen Z non/temporary subscriber, and Millenial “bandwagon” subscriber. These personas and their needs are detailed in the image below:
Jobs to Be Done extracted from personas:
Betty (Gen X, long-time CR subscriber and loyalist who believes in and supports CR’s mission and values):
- When my WiFi suddenly stops working and I’m not tech savvy, I want someone to fix the issue for me, so I can quickly resume watching my YouTube videos.
- When my oven suddenly stops working while I’m cooking and I’m not tech savvy, I want clear instructions on what my next steps should be, so I can bake dessert for my grandchildrem.
Trevor (Gen Z temporary CR subscriber, unfamiliar to CR’s mission and values):
- When traveling to a new place with my shiny new camera, I want to know if my existing warranty covers international damage, so I can decide if I should invest in a protective case for my camera.
- When my student Amazon benefit discount has ended, I want to evaluate whether it’s worth resubscribing to Prime, so I can pay the least amount possible while still enjoying the benefits of shipping and more.
Dylan (Millenial, new home-owner and expecting parent, recent CR subscriber who is familiar with CR’s values and mission):
- When I’ve ordered a new crib and some screws are missing, I want to return and replace the crib with minimal hassle, so I can ensure I have a crib in time before the baby arrives.
- When I notice that Netflix has increased its monthly price without my knowledge, I want to know if I can dispute the charge, so I can secure a refund, prevent future instances of being taken advantage of by Netflix, and understand the necessary steps to take if this situation occurs again.
Design Fiction
After defining the JTBD for each persona, we wrote 5 pieces of design fiction, each elaborating on one of the scenarios listed in the JTBD above and describing how the agent would act in the scenarios to help the persona that is facing the issue. The 5 design fictions can be found here.
This exercise allowed us to ideate more concretely how the agent would solve the needs of the different personas we defined.
Each of us worked on one piece of design fiction. After reading each others’ fictions, we highlighted agent actions that we found interesting and intriguing. The specifics of how the agent conducts different tasks can be found in our design fictions (the highlighted parts).
Moving Forward…
While we were discussing the different tasks that the agent could conduct and how it would do so, we quickly realized that we were being too ambitious. An agent that acts on all three phases (inform, guide, act) and conducts multiple different tasks within each phase will be an all-encompassing, omnipotent agent that will be difficult to design, develop and execute.
As a result, we concluded that we should now converge further and narrow down to a few tasks that our agent would conduct in any one phase of the consumer’s post-purchase journey. Our goal for this week is to now select and define those 1–2 specific tasks that the agent will conduct. We plan to do so using the pain points we identified from our primary research, the JTBD we defined in this sprint, and our understanding of CR’s overall values, mission and future direction.
We aim to design and test a lo-fi prototype by the end of next week, which will inform a mid-fi prototype that we will test the following week and present in our spring presentation.
(note: The work and knowledge gained from this project are only intended to be applicable to the company and context involved and there is no suggestion or indication that it may be useful or applicable to others. This project was conducted for educational purposes and is not intended to contribute to generalizable knowledge.)
