Product Discovery

Phil Hewinson
37 min readAug 5, 2020


My core focus since mid-May has been on talking to users and customers to better understand their motivations, behaviours and pain points. And then synthesising all of this information to figure out what products and features we should build next. This process is called “Product Discovery”.

As I mentioned in my last blog post, I’ve not conducted a single formal user interview before. So I’ve had a lot to learn and a lot to gain from this process. And I’ve had amazing support from both the Forward Partners Studio team, as well as LiveWork Studio, a service-design consultancy provided as part of Nesta’s CareerTech programme.

It’s been a busy and productive 10 weeks and we’ll soon be ready to get building again. But now that we understand our user’s and potential customer’s needs so much better, we’ll have much more confidence that what we’re building will create value for them. Here are some screenshots of the final prototype we produced (although I’m keeping some of the best bits secret for now, as it’ll be more fun to show you once we’ve built them):

And I’ve just hired my first employee who starts on Monday! So in two months, I’ve gone from being part-time to two full-time people armed with a ton of really valuable insights!

So let’s rewind back to mid-May and I’ll talk you through our product discovery process and then who I’ve hired and how …


One of my favourite experts in the area of product management and leadership is Marty Cagan, the founder of Silicon Valley Product Group. In addition to the book he wrote called Inspired (probably the best book for product managers and leaders at tech companies), he writes a blog. He wrote an article on the Origin of Product Discovery in June, which builds on his original blog post on Product Discovery in 2007.

These articles give some great insights into what the art of product discovery involves, why it’s important and how many startup founders fail because of their approach to this. He compares it to the process that pharmaceutical companies take to discover new drugs, and the high-risk nature of that process. Similarly, building software products that people love, which solve real user needs in a usable, feasible and viable way is really hard. It’s worth pausing to acknowledge that.

Most problems that are worth solving are very hard to solve. There is a long, hard road ahead for us to find “product-market fit”, which is the successful outcome of product discovery that we and every founder is aiming to achieve. And that’s the only thing that matters. Until we’ve built a product that people love, that’s all we should be focusing on. For us, that means building a product that helps people discover and get into their dream career. It also means finding a way of generating revenue, and scaling that profitably so that it’s viable.

My primary goal for the foreseeable future is exactly that, until we achieve it. And I intend to spend the least amount of time that I have to on anything else.


During the early days of the lockdown, I took the opportunity to read a few books on product discovery.

The Mom Test

The most popular book that everyone seems to be reading these days is a short book called The Mom Test. It talks you through how to talk to users and customers. The most common advice from respected advisors in the startup community is to talk to your users. And I think that’s because startup founders are often coders who like to build products. But they often find it harder to go and talk to people.

But once you do talk to people, what do you talk about? It’s very easy to introduce bias into the conversations you’re having in the way you ask questions. Or even worse, you might end up selling your idea to whoever you’re talking to. Instead, it’s important to try to get unbiased insights that will help you learn about them. And the best way to do this is to focus on their past behaviours. And then dig into their motivations and pain points. This book helps you prepare for these conversations, so you learn helpful insights that get you on the right track, rather than the wrong one.

Lean B2B

Lean B2B is a book that focuses specifically on conversations you might have with other businesses who are potential future customers. These conversations are often quite different to those you would have with consumers. So it talks you through how to prepare for these and set them up. And then what goals you should have with them and how you should approach the conversation.

When Coffee & Kale Compete

When Coffee & Kale Compete is a book that introduces a product discovery technique called “Jobs to be Done”. And the book is available as a free PDF 😊

There are many other product discovery techniques, such as MVP (that Eric Reis talks about in his book The Lean Startup), design thinking (more on that next) and customer development (which is partially covered in Lean B2B above). All of these discovery techniques have their place, and it’s often powerful to use a combination. I talked about the considerations for choosing a design thinking approach vs. an MVP approach in my last blog post. Going forwards, I’ll be using a combination of these approaches as and when they make sense.

Jobs to be Done is based on the premise that customers don’t have needs. Instead they just want to make progress within the system they belong to. It’s the idea that customers struggle when they try to make their life better, but they don’t know how. This struggle is called a “Job to be Done”. So in our case, getting into a new career that you’ll enjoy more is a “job to be done” for some people. And within this, there may be more granular jobs to be done, as we learnt from our user interviews. In fact, this is one way that we synthesised what we learnt from our user interviews (more on that later).

This book really helps to build an understanding of this product discovery technique and how it can be applied.

The Field Guide to Human-Centered Design

The last book I read was called The Field Guide to Human-Centered Design. This was written by, a leading design and innovation company. And it’s also available as a free PDF!

A human-centered approach to product discovery is a design-led approach whereby you start by interviewing users. You then slowly build up insights using a series of design techniques. And then by putting prototypes in front of people, of increasing fidelity, to see how they respond. The principle here is to de-risk delivery as much as possible, so that what the software engineers end up building is more likely to work. But it’s important to strike the right balance between this and actually building something. Because you don’t really know if someone will use a product until you actually give it to them.

This book is crammed full of useful techniques and approaches to draw insights from people during a conversation with them. It also talks through the theory and psychology underpinning it all. One of my favourite techniques, that we used a lot during our user interviews, is card sorting. One example of how we used this was when we showed a user a list of features we were considering. We asked them to sort the features in order of which they think would be most helpful. While they did this exercise, we asked them why they were placing them in this order. This let us dig into their motivations and identify which pain points (or job to be done) the feature was addressing. We used the virtual whiteboard, Miro, to do this exercise with participants remotely.

User Interviews

So in mid-May we started our process of talking to users.

Our goal with the user interviews was to talk with 10–15 people who are actively seeking to start a new career. As that’s what we’re here to help with. We wanted to learn about what they’ve done so far to try and change careers. And we really wanted to understand the areas that they’ve struggled with the most. From these conversations we could then learn more about our target users so we can figure out the best ways of helping them.


The first question was who to talk to. Besides finding people who are actively seeking a new career, we decided to first talk to people who broadly fit within what Nesta calls the National Retraining Scheme (NRS) cohort. These are the 8 million working adults in England who don’t have a degree. They also earn less than £35,000 per year and are employed or furloughed in an industry or role at risk of automation. As a finalist in Nesta’s CareerTech prize, we need to build a solution that helps this group of people. So it made sense to start our research by talking to these people.

How To Recruit?

The next question was how we should go about finding and recruiting these people so we can talk to them. We decided to try using our existing product at to do this. Just before showing the final results for a user who answers the 100 survey questions, we asked them to tell us if they are actively looking to start a new career. We also asked them what their highest level of qualification is. We recruited users using Facebook ads, targeted at people aged 25–44 who live in the UK. This wasn’t perfectly aligned with how Nesta defines the NRS cohort. But it was the best we could do in an easy way and I think it was close enough.

For those users who fit within our criteria, we asked if they could spare an hour to chat with us. And in return, we’d give them a £20 Amazon voucher. We then enabled them to book a slot with us using Calendly, which integrated with our calendars so people could easily find and book an available slot. On completing their booking, Calendly redirected the user back to our site so they had a “thank you” message. They could then see their final results.

Recruiting our First Users

So I built this flow in a couple of days and then switched on our Facebook ad campaign to recruit users over the weekend. After three days, I only managed to recruit two users for an interview. And I had spent £100 on ads to acquire them, costing £50 per person.

So I doubled the incentive from £20 to £40. I also forced the question on them before they saw their final results (at first it was an optional step as part of their final results). And I tripled our ad spend from £30 to £100 per day. That did the trick! Literally overnight we recruited another 3 participants to fill up our interview slots for Thursday 🎉

Preparing for our First Interviews

We spent some time crafting a discussion guide to help us with these interviews. We started by stating our goal, which was two-fold:

  1. Identify how we might be able to help users change into the career they want to pursue
  2. Identify if there are any products or services that the user might be willing to pay for

We then stated the top questions we wanted answering and our best guess as to those answers. The questions aligned with our goals, and focused on their motivations and challenges for changing careers, as well as whether they have spent any money.

We wrote an introduction for how we would start each call. And then we created different sections that started by building context around their current career. And then digging into their motivations and behaviours for changing careers. We had separate sections for the two main parts we were interested in, which was discovering and getting into a new career. We also prepared a couple of exercises that we did with them using an online virtual whiteboard called Miro. These exercises asked them to sort which potential features they would find most helpful and what they value the most in a career.

Our goal was to spend no more than an hour with each participant. So we aimed for this in terms of the number of questions in our guide. But we had bonus questions if we had time. These focused on their approach to writing CVs, job applications and preparing for interviews.

Interviewing our First Users

The Forward Partners Studio team ran these interviews as they have a lot more experience in doing these than me. Given the lockdown situation, the only option was to do these remotely. So we used Zoom to have video calls with each person. This conveniently gave us the opportunity to record the calls (with the person’s consent of course), which allowed me to watch each interview myself.

Only three of our first five interviewees turned up. But they went really well and we gained a lot of helpful insights. We then ran a further six interviews the following week, and I ran one myself too.

Broadening our User Base

As we were hitting our stride with interviews, we decided to take the opportunity to interview a broader range of users. We started by interviewing people with degrees. And then we interviewed younger people aged between 16 and 24.

All in all, we interviewed 23 people over 3 weeks:

  • Ten people from the NRS cohort
  • Seven with degrees
  • Six aged 16–24

One of the challenges we had was people turning up to their scheduled conversations. Only 1 in 3 people turned up on average. We thought it might be because they weren’t set up to use Zoom. So we gave people instructions to install Zoom when they signed up. We also thought they just forgot about it. So we sent automated and manual reminders over email and text. It improved the turn-up rates slightly, but not by much. Interestingly, the higher a person’s educational level, the more likely they were to turn up. Although our sample size was quite small, so I’m not sure how statistically significant this finding was.

Customer Interviews

While the Forward Partners team were busy interviewing users, I started the process of setting up and running customer interviews.

My primary goal here was to see if I could find a viable business model. I.e. a way of making money so we could build a product offering that was sustainable and scalable. I was interested in talking with three main customer types:

  • Recruiters who might pay for access to talent
  • Training providers who might pay for new students
  • Career professionals who can provide helpful context on all areas

A Failed Attempt to Recruit our First Customers

We’d obviously need a different approach to recruiting customers vs. how we recruited users. We couldn’t really use our existing product. So we’d have to try other channels like LinkedIn and our networks. And incentivising people’s time may not work quite the same. So we’d have to think of other ways of persuading people to take time to chat with us.

My first attempt was to use LinkedIn. I decided to target recruiters of Care Workers, as that was an in-demand job at this point during the pandemic. So I signed up for a free trial of LinkedIn premium’s offering, so I could do more advanced searching. That also let me send InMail messages directly into people’s LinkedIn inboxes.

I spent a full day researching the top companies who recruit Care Workers. I did this by looking at which companies posted the most number of Care Worker job ads on job boards like Adzuna and Indeed. And then I used LinkedIn to find relatively junior recruiters at these companies. Of the 48 people I shortlisted, I sent messages to 15 of them. I spent time carefully crafting a message to try and persuade them to take 45 minutes to talk to me, offering some compensation for their time.

A few days later I only had one reply. And that was from a lady who said she didn’t have time to talk. So back to the drawing board!

Recruiting Customers through my Network

I then sent emails to a few friends who work or have previously worked in recruitment. And I also sent an email to the Forward Partners team to see if they could help.

Jasel, Forward Partners’ Head of Studio, posted a request on LinkedIn for help. He had about 8 responses from recruiters in his network who all offered to chat! Result! So over the next week or so I started having my first customer interviews.

I reflected on why it worked so well to go through our own networks. To have someone take time out for you requires some exchange. For our user interviews, we were giving them money — a £40 Amazon voucher in exchange for an hour of their time. For these customer interviews, we were essentially trading our own social capital we had built up in the relationships we’d formed with them over time. People who know and like you will do you favours if you ask. J must have a lot of social capital and strong relationships to get that response rate!

As I had done with the user interviews, I spent some time crafting a discussion guide for each customer type (recruiters, training providers and career professionals). This would let me focus on extracting the most important insights.

30 Interviews Later! …

Nat, Forward Partners’ Head of People, also helped tremendously in setting up more recruiter interviews. She’s well connected in the recruiter space and sent a request to her network. She apparently had an avalanche of offers to help. So I spoke to a few of her contacts. I also spoke to a few people from my direct network. All-in-all, I spoke with 17 recruiters. And after a few conversations, it became clear that I should talk with recruiters who hire for entry-level positions, as those are the roles that are relevant for our target users who are trying to start a new career. So more of my latter recruiter conversations were with those.

I also spoke with six people who work in the training market. These were mostly independent training providers. I quickly learnt that the opportunity to make money here wasn’t as strong as it was with recruiters.

And finally I spoke with seven people who work as career professionals in some form or another. They provided really helpful context on many different areas that were new to me.

Over my 30 interviews, I spoke with people from:

Synthesising Information and Forming Insights

So by the end of June, we had spoken with 53 people! Which was a lot of information! So, what next!?

Well, now we had to try and synthesise this information and form insights. The goal was to see if we could identify patterns across the interviews. This would form our core insights that we could then use in our product planning.

Synthesising the User Interviews

For the user interviews, I made structured notes as I watched each video to record:

  • Their current career
  • Motivations for changing
  • What careers they are considering
  • Challenges they’ve faced and what they’ve done about those
  • Whether they have spent any money
  • What they’d value the most

But that still surfaced a lot of information. How should I go about drawing insights from all of these notes?

Applying Service Design Methods

Being a finalist in Nesta’s CareerTech prize helped a lot here. Part of their programme of activities were sessions on service design led by a consultancy called LiveWork. They also provided monthly 1:1 coaching sessions. This gave me the opportunity to discuss methods I could use to synthesise all of this information.

The first method I employed was to map all of the user needs I identified into the different stages of a user’s journey from career discovery to getting a job. I did this by recording each user need in a separate cell in a Google Spreadsheet on the column corresponding to that stage of the user journey. I then grouped all similar ones together. This gave me an indication of how many people had the same needs, which surfaced the greatest needs that people faced.

I also colour-coded the cells based on which cohort the person who expressed the need fell into. The different cohorts were 16–24 year-olds, 25+ without a degree or 25+ with a degree. This yielded further helpful insights that were unique to the NRS cohort (25+ without a degree).

One of LiveWork’s suggestions was to build a value exchange network, showing the relationship between different people in the careers space. So I produced this:

They also suggested producing a journey and opportunity map, to consolidate the insights from the user and recruiter interviews into a sequential journey. And then highlight the opportunities at each stage. So I produced this:

Our Core User Insights

People’s main motivations for changing careers was to find a job they enjoy more, and also to grow and progress. And they struggled with all stages of the process.

People who struggled to identify the right career to pursue often talked to people they knew to explore further. They were particularly attracted by work experience opportunities to try jobs out. For people who struggled to find a way into a career, they’re often overwhelmed by the volume of generalised information available. Instead they want clear, personalised steps to follow that would guarantee them a job.

People struggled to identify open roles that matched their current skills, experiences and qualifications. When it comes to applying for jobs, the biggest frustration people had was not hearing back from employers after investing time in a job application. And also dealing with a lot of rejections.

People in the NRS cohort (without a degree) had some unique struggles. These included having certain constraints around pursuing different careers. Such as their physical health, having better hours than their current job gives them, and identifying more realistic options. They were uniquely interested in unpaid work experience opportunities. And they demonstrated a greater struggle around finding suitable open roles based on their skills, experience and qualifications.

Young people were unique in that they were more confident in the career they wanted to pursue. They were also more optimistic (and probably less realistic). They were naturally more focused on qualifications and education, rather than skills and experience. And they expressed more interest in support with interviews and writing a CV, as this was new to many of them.

Our Core Entry-Level Recruiter Insights

Recruiters of entry-level roles get hundreds of applications for each role they post on a job board. And hiring for entry-level roles has slowed right down during the pandemic. Which means that there are fewer potential customers who are looking for talent. But those who are looking face bigger pain points around screening candidates, as they are getting a lot more applications than normal.

Employers often have specific requirements they look for when screening candidates. Such as age, qualifications, specific universities and past experience. The main criteria recruiters test for in candidates for entry-level roles are motivation, their personality and their attitude. Diversity is an important motivator for companies when hiring for entry-level roles. This is because hiring for diversity at the entry-level has proven to be the best way of building a diverse workforce in the future.

Most recruitment for entry-level roles is done in-house. And in-house recruiters are more interested in sourcing great talent rather than having end-to-end support on the whole recruitment process, which is what a recruitment agency typically provides.

Forward Partners’ Synthesis

The Forward Partners team independently synthesised their insights too from the user interviews. This helped to create a more complete picture of people’s needs. They did this in a slightly more granular way, which added value.

They used a service called Otter, which transcribes each video recording. This enabled them to more quickly parse the interviews and copy the key things that people said. It was incredibly cheap too. It only cost about £12 to transcribe about 20 hours. Whereas manual services would have cost about £1,000.

The Forward Partners team then placed key insights into separate cells in a Google Spreadsheet in one of three columns for each user:

  • Job to be done (as I talked about earlier)
  • Quotes
  • Feelings

We then used this directly in our week-long design sprint …

Design Sprint

In mid-July we ran a week-long design sprint. All of the work on conducting the user and customer interviews and synthesising everything we learnt fed into the sprint.

The goal of the sprint was to answer this question:

How might we help people get into a career they want to pursue and also create value for recruiters hiring for those entry-level positions?

The output of the sprint would then give us clarity on what direction to take with the product. It would also guide us towards the next set of features we should build and test. The sprint was modelled after the book Sprint by Jake Knapp, which he invented during his time at Google.

Due to the pandemic, we ran the sprint remotely over Zoom. Eight people were involved, including six people from the Forward Partners team. A career expert also joined us, who we recruited externally. Dharmesh, Forward Partners’ Head of Product facilitated the sprint. He did an excellent job in managing our energy and guiding us through what was an intense and complex process.

We used a tool called Mural, which is a virtual whiteboard. It’s similar to Miro that I mentioned earlier, but richer in features. Very conveniently, they have a template modelled after the book Sprint. We used this directly, and it really helped to structure all of our notes over the week.

Day 1 — Understand

Day 1 was all about context sharing to get everyone up to speed. We started with introductions and an ice breaker. I then spent some time talking about the company, the vision, where we’ve come from and where we’re going.

Identifying the Biggest Risks

We then did an exercise of writing down all of the questions we each had and what we saw were the biggest risks for the product and company. We wrote these down on virtual post-it notes directly on our Mural board.

Dharmesh then led a voting session on what we thought were the biggest risks. The top two risks that emerged were:

  1. Can we create value for recruiters?
  2. Are we able to help users create a viable pathway into their chosen career?

Sharing Insights and Mapping a User Journey

We then spent quite a bit of time going through all of our user interviews using the synthesis that Dharmesh and his team had done. We each looked at the insights from a few users and wrote post-it notes phrased as “How might we …” to capture thoughts around how we might solve the different needs that people have. We did this to spur ideas and creativity later on.

We then spent some time sorting through these 100+ post-it notes on our Mural board. And then grouping them into different categories corresponding to the different stages of a user’s journey.

We repeated the same exercise for the recruiter interviews.

We then worked independently on creating a single user journey as we understood it from start to finish. A couple of us worked on the recruiter journey instead. And then we worked collaboratively to bring these journeys into a single user and recruiter journey and how they intersected with each other. And finally, we highlighted the part of that journey that we chose to focus on for the duration of this sprint. The final journey that we chose to focus on, which is exclusively focused on the user, is:

Defining an Archetype

The last thing we did was to define the archetype that described the people we wanted to focus on helping. Archetypes are similar to Personas, but described a broader group of people rather than a particular type of person. We decided to focus on career changers (employed adults aged 25–44). As that overlapped with the NRS cohort that Nesta wants us to focus on. And we also felt that they expressed a stronger struggle when we interviewed them. We spent some time digging into their motivations, behaviours and frustrations.

Day 2 — Sketch

Day 2 was all about brainstorming and ideating.

We all had homework to do beforehand though. We each had to prepare a “lightning talk”, sharing a few different products we’ve each come across that help people with their careers. Or products that have interesting user experiences that could inspire us.

We then spent the morning presenting these talks to each other. They were brilliant! We must have covered 20–30 products. We covered products that cater for every part of the career journey a user is on. They also included dating apps and diet products that had interesting and inspiring user experiences. These really helped get our creative juices flowing for the afternoon …

After lunch, we ran an exercise called “Crazy 8s”. In preparation, we each had about half an hour to gather all of our thoughts and ideas. We then had about 20 minutes to sketch 8 ideas on a piece of paper for different ways we could solve the user needs as we understood them. Normally, you only have 8 minutes to do this exercise (1 minute per sketch). But we didn’t feel that time constraint was helpful to us, so we took a bit longer.

Between the 8 of us, we yielded about 60 ideas! We then spent 30 minutes presenting back our ideas to the group. This gave us all a lot more ideas. So we then did another round of Crazy 8s to yield another 60 ideas! And then we presented them all back again. All-in-all, we generated about 120 ideas. And I recorded the presentation of our ideas, which gave me the opportunity to watch them back again a few days later (and for future employees to watch them too).

Day 3 — Decide

On Wednesday, the goal was to go into detail on a single user flow so that we’d be in a position to prototype it.

We started the day by voting on the 120 Crazy 8 ideas we had generated the day before. We then spent over 2 hours discussing why we voted on each one and then talking through the ideas together. I also recorded this session. This gave me the opportunity to watch back over this session a few days later.

This fed into an activity where we each created a storyboard of what we think we should test with our users to answer our questions. The idea here was to bring together all the strongest ideas into a single user flow that we could then prototype to see how users reacted to it. We each worked on this independently. And then we each voted on which one we thought was best.

We then used the top voted flow and complemented it with a couple of missing parts that were in the other flows. This formed our master flow for our final prototype.

We then spent a couple of hours going deep into each screen of this flow. We broke off into groups of two (using Zoom’s cool breakouts feature) and took a screen at a time. Each breakout had a 20 minute time limit to brainstorm all of the information that might be needed on that screen. We started with the riskiest and toughest screens, so that we’d have time to cover the hardest parts.

The end result of day 3 was a final flow of screens. Along with details of what should be included on each screen.

Day 4 — Prototype

We all had the day off on Thursday, except for our designer Josh. The goal for today was for Josh to produce a high fidelity prototype that we could use to test with users on Friday.

Josh produced a prototype in Figma, which is an amazing web-based design tool. Essentially, he produced a series of screens (images) with clickable elements that went to the next screen. So people could view each screen and click on certain buttons so they can navigate through the flow. This would help people feel what the product would be like, so we could get their first impressions before going to the effort of building it.

Dharmesh and I had a call with Josh to talk through initial wireframes he had produced. He then did an incredible job of putting the whole prototype together by the end of the day.

It encapsulated all of our big ideas that came out of the design sprint, in a single user flow. It looked and felt like a real thing, which was exciting! Although it felt a little daunting because I knew that to build it all would take well over a year. But the goal wasn’t to then build it, but instead to see how people reacted to the different parts. We’d then know which features to focus on in our next steps.

Day 5 — Test

On our final day of the design sprint, we tested our prototype with five users (we actually did these over Friday and Monday). During the week, we switched on our Facebook ads to recruit 15 users ready to test with us at the end of the sprint. We knew from scheduling previous user interviews, that only 1 in 3 people actually turn up to the interview. So we scheduled three times as many interviews that we needed. As expected, only five people turned up, which gave us the insights we needed.

Building Towards a Minimal Viable Product (MVP)

Synthesising our Learnings and Ideas

I then spent some time digesting the feedback and insights we gleaned from showing users our prototypes. I summarised the key features that stood out for people. And also some of the emotions that people expressed.

I also re-watched the recordings of our Crazy 8s and our discussion around them. I then drafted a product plan of features we should consider building in the coming weeks and months, and prioritised each one. The top priority features are ones we’d probably need to build in order to have a minimal viable product (a product that people could use and get value from towards our goal of helping them discover and get into a career they’d enjoy).

Dharmesh also reviewed the interviews and recorded his insights as well as key hypotheses we should test. We then spoke together to form a plan for how we would decide which features to build and then design our MVP.

The Next Two Weeks

Over the next couple of weeks, we’ll work with Dharmesh and Josh on storyboarding what the flow will be for our MVP. We’ll then go into the detail we need for each screen (as we did for the design sprint prototype). And Josh will then design it. Then we’ll be ready to build it!

Hiring My First Employee

My second big priority for the last 10 weeks has been to hire my first employee!

I talked about hiring two people in my last blog post. A senior software engineer and a researcher / operations person. On reflection, I decided not to hire a software engineer for now. I’m able to do the coding myself (and that’s my favourite part!). And hiring someone will increase our burn-rate (the speed at which we are spending money). I’ve decided to wait until we are generating revenue before hiring a software engineer. Then we can maximise our runway (length of time we have until we run out of money). I think this will give us the best chance of success.

But I was keen to hire a non-technical employee. I decided to post the role as “Employee #1”. I felt that this sounded more exciting. And it better encapsulated the responsibilities of the role, which are incredibly broad and varied. I’ll spend most of my time coding, so they’ll need to pick up most other things! Initially, the main focus will be researching how we can help people get into a small selection of careers. And then the focus will likely be more operational, and running the programs we experiment with. And eventually scaling them.

Writing a Job Spec

So I started by writing a job spec. I took some time to really sell the company, the mission and the opportunity. I think we have a really compelling social mission. And this role is an amazing opportunity to learn, grow and be part of something exciting at a really fun stage. So I wanted to communicate that.

I then took a stab at writing the “requirements”. But this was hard because I actually don’t know what their main focus will be 3–6 months from now. So how do I decide who I’m looking for?

For the job spec, I decided to share some insights into the types of things they’ll be doing in the first couple months. Such as research and project management. But then I shared that this will be very varied beyond that, and that there are tremendous growth opportunities as the company grows.

And that was that really. I didn’t want to be too prescriptive in case we missed the ideal candidate.

What We Were Really Looking For

But what were we really looking for?

Well, I decided to primarily look for three attributes:

  • Intelligence: Given the varied nature of the role, I need someone who can learn really quickly. Then they can pick up whatever is required of them
  • Motivation: I want someone who is really excited by the social mission. Then they’ll have a lot of energy for the role. And they’ll be more likely to stick around for a long time
  • Hard Working: I need someone who will get stuff done and work really hard. Then we can keep moving fast and give ourselves the best possible chance of success

I also cared about their ability to deal with tremendous uncertainty, their product sense, how well they communicate, how important learning and growth is to them and whether they are a good culture fit (i.e. do I get along well with them).

Side Note About Job Specs

I found this process of writing a job spec as an employer really interesting. Especially as what I wrote and what I was actually testing for were quite different. And I had good reasons to do it this way.

But it makes me wonder about all the other job specs out there and how accurately they really reflect what employers are looking for. This is the core part of the problem we’re looking to solve next in our mission. And I think the best way for us to help people get into careers is not to parse all the job specs. But rather to talk directly to the decision makers who are hiring for those roles.

Recruitment Process

The recruitment process served two purposes for me. Obviously I wanted to hire an amazing first employee. But I also wanted to learn as much as I could about recruitment as that’s aligned with our company’s mission.

I employed the services of Forward Partners’ people team to help with recruitment. And they provided an amazing recruiter called Suf who led the charge.

Posting the Job Ads and Managing Applications

Suf put my job spec on Workable, which is an applicant tracking system (ATS) they use. ATS’ are nice pieces of software that serve a few functions:

  • Craft a job ad. This consists of a description and requirements, as well as an application form that you can customise
  • Easily post your job ads on all the job boards you want to. Such as Indeed and Reed (there are many, some free and some paid)
  • Manage inbound applications. This includes sending messages to candidates and moving them through the process of screening, interviewing and final job offers. It can also capture comments from each interviewer against the candidates’ profiles

There are a few other ATS’ that different companies use. Such as Lever, Greenhouse, Taleo and Workday, plus lots of smaller ones too. We’ll need to think about if and how our product integrates into these.

Learning about Other Products on the Market

There were some places that Workable wasn’t able to post our job ad. So we had to do so manually. One of these was Otta, which is a relatively new startup focused on jobs in tech in London. They have a great user experience and are in a similar space to us. So I was keen to post there. Both for the relevance of candidates, and to get a closer look at their offering. One of the things they do well are nice-to-read profiles of companies. Here’s our profile they put together for us.

I was keen to understand more about how Tempo works. This is a startup with a mission of making recruitment fun. Candidates record short videos, which makes it easier and more enjoyable for companies to browse candidates rather than reading dry CVs. They also have an algorithmically generated feed of top candidates that match your requirements. As my role was so unique, it didn’t work so well for me. So they manually intervened to populate my feed of suitable candidates.

I was keen to try out Give a Grad a Go too. Their process is a lot more manual and they don’t really have a software product to use. I completed a form and they got in touch to have a call. They then sent through a few anonymised candidate CVs over email for me to review.

200 Applications in a Week!

We received over 200 applications in a week! As so many people have been furloughed and made redundant, there are a lot more job seekers in the market. So jobs are getting far more applicants than they’d normally get. This is bad news for candidates as they have far more competition and it’s much harder to stand out. But it’s good news for employers, as we can probably hire better calibre people than we’d normally be able to. Although it’s more work to screen a lot more applications!

I’d say the best source of quality candidates came from the job board Escape the City. This website is geared towards professionals who work in finance and consultancy and want to change careers. Their main focus are careers in tech startups. Normally employers have to pay a few hundred pounds to have their jobs listed and promoted. But for some reason our role was listed and promoted at the top of their listings for the whole week! Maybe they thought it was interesting, and decided to promote it for free?

Screening & First Interviews

I started screening all applications myself, scoring them against a scorecard. I devised a scorecard based on the eight attributes I was looking for (as I talked about earlier). And then I chose a subset of attributes that we could get some reasonable signal on at the application stage. I also asked some specific questions as part of the application process that gave us signals on their motivation, product sense and intelligence.

But after screening 50 candidates, I found it to be very time-consuming. So Suf screened the rest and gave me a shortlist of candidates to consider. I then reviewed each of their applications in detail and decided which ones to proceed with.

I then had first-round phone interviews with about 10 candidates. Before the interviews, I spent some time crafting an interview guide. Then I could ask them questions that gave me some signal on key attributes I was looking for. I customised some of the questions based on their experiences. And I used a new scorecard for this stage of the process.

Task & Final Round Interviews

We advanced a handful of candidates to the next stage, which was when we asked them to complete a task. For me, this was the most important part of the process and when we were likely to get the strongest signals on the candidates. I asked each of them to spend 2–4 hours researching how to get into a career as a Care Worker. The expectation I set was that it should involve some desk research and talking to a couple of people. I asked them to put together a short presentation, showing their approach, their results and product ideas they had for how we could use that research. We then asked them each to present this back to us as part of the final round interviews.

This task gave us a really strong signal on their:

  • Motivation and work ethic (how hard they worked on it)
  • Intelligence (based on their approach and synthesis of information)
  • Communication (the construction of their presentation and its delivery)
  • Product sense (translating their research into product ideas)

I also employed the service of Forward Partners to help do some final round interviews too. That ensured that I had some diversity of opinion when making a final decision.

And Our Employee #1 Is …

We had some really strong candidates come through the process, so it was hard to turn them down. Especially those who got to the final round interviews. But there was one candidate who really stood out. Her name is Emma Rosen.

Emma has a fascinating background. She started her career on the civil service fast-track programme. But after a year she found that she wasn’t enjoying her job at all. So she quit and tried 25 different jobs in a year. But not just any 25. She picked the jobs she’s always dreamed of doing. From helping out on an Alpaca farm in Cornwall, to photography at a wedding in Ibiza. And even archaeology in Transylvania!

Then she wrote about her experiences on a blog, all captured on her website at 25Before25. She then did a ted talk on career happiness. And she wrote a book on career happiness called the Radical Sabbatical (which I read last week — it’s great!). She founded a startup focused on helping people get work experience opportunities. And she’s worked on a bunch of other interesting projects.

So I’m delighted that Emma chose to join our mission here at Would You Rather Be!

From drafting the initial job spec on June 24th, we had an offer accepted from our top candidate on July 24th. Exactly one month later! I was pleased with how fast we were able to move. But I’m even more pleased that we found such a strong candidate, who is such a good fit for Would You Rather Be.

What’s Next …


My main priority over the next couple of weeks is to onboard Emma into the company as well and efficiently as possible. I’ve already invested quite a bit of time into this. I wrote a 15-page onboarding document. This covers the tools we’ll be using and all the context and linked resources that I could think of. Plenty of reading for August! I sent her my top books that are most relevant, on product and careers. I’ve organised all of our company’s documents using Google’s Shared Drives, and carefully permissioned things. And I’ve given her a new Google account (with a new email address), access to all the documents she needs, a paid Zoom account and set up a new Slack workspace.

Plus there were a lot of people-related things we needed to do. Such as drafting a new employment contract. And figuring out our People policies and procedures we need and do. Plus things like “right to work” checks, which are legal requirements. Nat in the People team at Forward Partners really helped with a lot of this!

And I’ll spend quite a bit of time with Emma over the next two weeks, sharing further context and working through our product strategy and plans together.

All The Small Things

My goal is to free up as much of our time going forward as we can to focus on product-market fit. Specifically, to build a product people love that helps them discover and get into their dream career. And then to make money, and scale it profitably. Until we do that, no other priority comes close.

But there are a bunch of things I still need to do. So my goal for the next couple of weeks is to try and tie off as many of these things as possible. I’m then on holiday for the last two weeks of August so I can fully recharge. And then in September we should be able to move really quickly on product execution without too many distractions.

So what are all these other things that I’m trying to tie off? …

Storyboarding and Designing our MVP

We still need to map our high-level prototype of all of our big ideas, along with our user insights and product ideas, towards an MVP. The next step here will be to storyboard our user flow for our MVP with Josh and Dharmesh at Forward Partners. We’ll then need to flesh out the details of each screen. And Josh will put a prototype together that we can use to build the actual product in September.

Researching Minimum Qualification and Experience Levels For All Careers

Part of the product experience we need to build is a user journey from the current final results screen (which shows 16 careers on average to someone) to one career that a person chooses to pursue.

One approach we need to take is to do some high-level filtering of careers. This will be based on someone’s level of qualifications (or aspired level of qualifications) and their experiences. For example, there is no point in showing careers that require a degree to someone who has no aspirations to get a degree. This body of work requires researching the high-level entry requirements for each career. I’ve started this work and I estimate it will take about 3 more full days to complete.

Researching How to Get Into Five Target Careers

In order to support people into a career they want to pursue, we’ll narrowly focus on just a few target careers. We’ll carefully decide which ones to focus on based on:

  • The demand in the market today (i.e. the number of open roles)
  • The popularity (based on data we already have from people completing our survey)
  • How much employers might pay for access to high quality talent (with is probably proxied on starting salary)
  • The ease in which someone can access a career (e.g. a career as a solicitor usually takes at least a year to complete a conversion course, so we’ll probably select careers that are more immediately accessible).

Once we’ve selected our five target careers, Emma’s primary task over the next month will be to deeply research how people can get into those careers. The primary method here will be to talk to hiring managers and recruiters for entry-level roles into those careers to see what they look for and how they decide on which candidates to hire.

This research will then feed directly into what product we build and how we can best support people to get into one of these careers (for those who want to).

Visual Design

I’m not a visual designer. I have some appreciation for good visual design, but I can’t produce it. As you can probably tell from how the website currently looks. So we’re employing Forward Partners’ services to help spice up the look-and-feel. They’ll help design a new logo as well as colour schemes, fonts etc. to use on the site. That’s happening next week.

Data Protection / GDPR Compliance

Data protection has been on my todo list for a few months. I have a privacy policy, but it’s one I cobbled together last year. So this needs a little work. I probably need a Terms of Use policy too. And a cookie policy. Plus I probably need some cookie opt-in that most sites have.

But data protection compliance goes much deeper. I also need to carefully consider:

  • What data I’m collecting, why and on what legal basis I’m doing that (and document that in my privacy policy)
  • Decide how long I should keep the data for and ensure I have policies and procedures in place to adhere to this.

This not only includes user data, but potentially data on employees, contractors, other potential customers and even job candidates.

I also need to review which third-party suppliers I am sharing my data with and ensure I have adequate Data Protection Agreements in place with them, and store these. These suppliers include:

  • MongoDB Atlas that I use for my database
  • Heroku that I use as my server
  • GSuite that I use for document storage
  • Slack that we’ve just started using for communication and collaboration.

And I owe a fee to the Information Commissioner’s Office.

We need to do this for our product and setup today. But also future-proof it so that we can execute rapidly on the product experiments we want to do, while staying compliant.

Nesta Midpoint Check In

And finally, a midpoint check in is due as part of Nesta’s CareerTech prize by the start of September. This includes:

  • A summary of our solution
  • Progress update
  • Planned activities and milestones
  • Learnings of our beneficiaries
  • Our use of labour market information
  • The potential market for our solution
  • How we plan to measure its impact

Thankfully, I already completed this over the past few days and I’ll submit it before I go on holiday in August. It was a lot of work, but was helpful to take stock of where we’ve got to, the work ahead and ensure we put important things in place. Such as measuring a baseline for our impact assessment.

Building Product

So the goal will be to start building product again in September. It’ll be a busy two weeks trying to wrap a lot of this stuff up. But then we’ll be free from distractions and able to focus come September.

I’m really excited by the speed we should be able to move in September though. With two of us working full-time. And armed with deep user and customer insights, some strong research and a solid prototype to build against.

I’m hoping we’ll make some really rapid progress and start helping people get into a career they’ll love!