There's a saying in the startup world that no business plan survives first contact with customers. And despite the promising results of the first test for Talent Compass, all assumptions about how such a service would work in practice needed validation.
Talent Compass is designed to be a career advisor/coach in the form of conversational UI aimed at job seekers from underrepresented and marginalised backgrounds.
A key decision in the plan for the Rehearse phase was to build partnerships with Employers and Intermediaries who could later provide job opportunities to Sophie, if Talent Compass proved a desirable product.
As a result, the primary focus was to validate Talent Compass as a desirable product for Sophie, while ensuring it was technically feasible to match Sophie with relevant open roles based on her passion and experience.
The best way to validate a service is to actually run a lean version of the service. Doing so requires a multidisciplinary team of digital experts able to navigate the ambiguity of testing a service that doesn’t yet exist.
For this, the Rehearse phase, Red Badger deployed a team with product (UX & Design), engineering, data analysis and delivery management capability for 12 weeks to iteratively test riskiest assumptions.
This approach ensures product-market fit ahead of creating a costly, production-grade service, while the short timeline keeps the team focused on fast feedback loops and aligned with the principle of metered funding.
Over the next few weeks, the ream relied on its capabilities to validate (or disprove) our assumptions on:
- Product desirability through interviews and testing with users and industry experts
- Technical feasibility through tech investigations and rapid prototyping
- Viability by working with Mission Beyond to execute the intended business model and shape milestones for scaling up the venture
Working backwards from a successful outcome
At the end of the Rinse phase, the team identified the riskiest business model assumptions to test in the first month. Assessing these would help determine whether to pivot or persevere for months two and three.
However, before diving into work, it was important to establish team processes and norms for how to rapidly test assumptions and maintain clarity of mission throughout.
To reduce ambiguity and stay aligned, the team agreed to leverage Red Badger’s usual product delivery process of working from a backlog. A pre-mortem was run to identify the biggest hurdles to overcome and define tasks for mitigating these hurdles. It also highlighted leading indicators of success to measure and refine across the 12 weeks.
Leveraging ‘Wizard of Oz’ testing to validate multiple assumptions
With the team setup and ready to go, it set to work planning a test to validate the following assumptions:
- A conversational tone and style is the most effective way to engage Sophie (product desirability)
- Sophie would value having her skills articulated in a way that makes it relevant to the job market (product desirability)
- It is possible to surface job relevant skills to Sophie based on the passions and life experiences she writes into the UI, with the use of a general purpose skills database (tech feasibility)
- It is possible to translate the skills surfaced by the database into relevant jobs to provide job guidance to Sophie (tech feasibility)
- The product’s branding lends it credibility, and users will be willing to engage with it (Viability)
As all of these assumptions are interlinked, the tech and product people came together to craft a Wizard of Oz test to validate them in the leanest way possible.
“To surface role recommendations for Sophie, we used a graph database since it's an area they excel at,” explains Carlos Roman, Tech Lead at Red Badger who headed up engineering on the project.
“Adopting the European Skills, Competences, Qualifications and Occupations (ESCO) to power our graph, we tested our assumptions more quickly. It helped us find gaps in the data and what it could or could not do for us.”
The database was exposed via a RESTful interface using Spring Boot. Out of the box, Spring Boot provides a way to create a simple domain model and RESTful interface to work with.
The first iteration of user testing involved running a script through the endpoint. Once the team was happy it worked, it was clear it could power a frontend.
“We put a GraphQL endpoint in front of the RESTful endpoint,” continues Carlos. “We did this to accelerate the work done by those working on the front end. As the front end was written in NodeJS using React, GraphQL made integrating the database with the web app easier.”
Working this way reduced the burden across the broader team since there was no demand on time to integrate platforms. Work was then split with half the team working on improving searches while the other half finished the frontend.
The prototype was then deployed to Render, where the team could showcase it to stakeholders. It provided a ready-to-use platform, free from concerns about infrastructure so the team could concentrate on interrogating assumptions.
Over the course of two weeks, the team conducted one hour sessions to test a lightweight version of Talent Compass with seven young people aligned with our persona, Sophie. After brief introductory interviews, each user was asked their initial impressions of Talent Compass while they viewed a static prototype of the home page.
The interview participants--all of which expressed interest in using the product at this point--were directed to the video conferencing chat, where Kai Compass (Talent Compass’ chatbot) asked a series of questions to uncover passions, skills and life experiences.
Behind the screen, the Red Badger tech team input search terms from this conversation into the skills database created. Once all questions had been answered and the skills database searched, a list of relevant skills and jobs was shared with the interview participant.
The outcome of this test was highly successful! Everyone that tested the prototype of the product was astounded at how accurate the skills and relevant jobs surfaced were (in comparison to other services used in the past).
As a result, the team proved the value of producing a higher-fidelity version of Talent Compass to test remaining assumptions.
Letting customer evidence guide focus
While user testing and interviews caused initial excitement about Talent Compass, a wider theme needed addressing. The team started to hear comments such as, “...it would be good to see a menu of [open] jobs and opportunities,” and “...will this let you add a paragraph to your CV? That would be cool”.
In other words, it wasn’t enough for Sophie to have transferable skills surfaced, she wanted help to enter the workforce. Surfacing live job opportunities was beyond the scope of the service the team initially considered – yet Sophie was clearly telling us the service needed to do more to provide real value.
This critical moment lead the team to turn attention to the platform that will hold the entire eco-system.
Focused initially on the product beneficiaries interact with, now the connections between the opportunity providers and the joining up of complex user journeys came into view.
A national digital infrastructure would be developed connecting the ecosystem together securely, robustly and at scale.
Working with the Mission Beyond team, a course was plotted for the platform, one that would enable the benefits to flow between all participants: corporates seeking diverse talent, intervention and impact providers, and diverse talent seeking opportunity.
“The Red Badger team’s flexibility and adaptability is remarkable. From their initial dedicated focus on ensuring every interaction with a beneficiary was valuable, to at-scale; platform requirements driven by the ecosystem of thousands of entities offering opportunities.” says John Godfrey.
“It’s a testament to the agile ways of working deployed, commitment to delivering the right product and the depth of talent within the team to move seamlessly between a single user need to a national digital infrastructure platform and keep value and impact at the heart of every decision.”
The team extended their perspective on the product and users.
18-25 year old seekers of opportunity are just one peice of the puzzle. In order to surface live job ads to Sophie, the team knew the needs of corporations and partner organisations were essential to understand.
To ensure Open Doors could serve all sides of the ecosystem, the team explored how the service could provide value to them, as well as Sophie.
The team mapped out how Sophie might interact with other segments through the Open Doors platform.
A customer value proposition was created and a reworked business model canvas drawn up to generate an assumption backlog and required tests to identify the desirability, feasibility and viability of the platform.
Addressing the mission
Within 5 weeks of the project start, the team set to work designing and validating an interconnected social mobility ecosystem. To keep things lean, the team tested a static prototype of a partner portal designed purely on assumptions. This prototype demonstrated how an employer could:
Understand the Open Doors proposition for employers and mentorship programmes:
Sign up to Open Doors by providing details about their organisation’s size, sector and available roles:
View potential candidates interested in their available roles (based on the job matching process designed for Sophie):
Review the profile and application of people like Sophie who have submitted an application:
See anonymised diversity data of applicants:
The team also expanded on the prototype for Sophie to include a list of open roles she could apply for. Tests were conducted with the same users that provided feedback on the initial product and further interviews arranged with large organisations.
Based on what the team found out, the team iterated on the designs to make job descriptions more user friendly.
This included features in the partner portal to articulate the Social Mobility USP and to ensure differentiation from existing recruitment platforms while enabling partner organisations to offer Sophie job openings and other development opportunities.
Finally, after all assumptions were tested and we iterated on Open Doors to a point both employer partners and users like Sophie were expressing interest in signing up, the team had something that could form the basis of a build backlog and product roadmap.
The final step was to equip the Mission Beyond board with the learnings from the Rehearse phase and support them in their quest to secure funding for the next stage: building the product.
A pitch deck was created to cover user testing, product vision, wireframes of both product and platform, infrastructure ideas, data security and analytics methodology.
“With such rigorous product development, user testing and analysis we knew we had a great service on our hands,” says Cain Ullah, Red Badger Board Member and Founder and former CEO.
“The supporting materials provided by the team alongside the consistent and in-depth feedback sessions throughout this project has made our job simple in communicating the tremendous value of this service.
"We are extremely hopeful of securing the necessary investment to bring this platform to life and start tackling the social mobility challenge.”
Graduating from product strategy to platform delivery
The platform can and will address a pressing need in today’s employment market - connecting job seekers looking for meaningful careers with large organisations seeking diverse talent.
Red Badger’s Product Strategy framework delivered both a product concept and empirical evidence of its desirability, feasibility and viability in 15 weeks. All while ensuring the highest chance of a successful launch.
And through this process, Mission Beyond now has a clear business case for future investment in a build phase, and a clear sense of the timeframe to get there.
The team can’t wait to see the impact this will have on young peoples’ careers and are confident it will change the conversation on social mobility in the months to come.