I recently raised £550k for my startup, Would You Rather Be!
This includes a £500,000 venture capital (VC) investment from Forward Partners and three angel investors. We’re also one of twenty proud finalists of Nesta’s CareerTech Challenge prize, who have awarded us a £50,000 grant.
I started this project last June to help people find a satisfying career. And I wrote a blog post at the end of October about the progress I’d made to-date. Since the new year, I focused on raising money so I could work on it full-time and build a small team of people to work with.
I closed the VC round at the end of March, received the Nesta grant at the end of April and have recently finished and released a new second version of the career discovery app at www.WouldYouRatherBe.com. Feel free to try it out — it’s much better!
In this blog post, I’ll share my journey on this project since my last update in October. I’ll cover how I carried out a bunch of failed experiments in trying to make money, how I raised £550k, the impact of Coronavirus, what I developed for version 2 and why, the foundations of the business that I built and where I plan to go next …
Failed Monetisation Experiments
My goal at the end of last year was to see if I could make money from the existing career discovery product.
Charging Users to See Results
I started by giving people 3 career suggestions in their final results for free, but charging them £10 to see a more complete list of 30 career suggestions. I implemented this in a couple of days using Stripe.
But it didn’t work! I tested with a few hundred people and not a single person paid to see more results. I then tried a more aggressive tactic by not showing any results for free, but charging the user £5 to see all 30 results. Again, not a single person paid. Moreover, I started getting negative feedback on my Facebook posts as people were annoyed to answer 100 questions to then have to pay to see their results. Which was fair as I didn’t warn them up front that there would be a charge.
So I quickly reverted back to a free app and replied to the angry comments on Facebook to apologise, saying that it is now free and that we were just trying an experiment.
Displaying Ads Next to Each Question
The third and final experiment I tried was placing ads next to every question. An old friend of mine who I worked with at Facebook said this could be viable, as each person would render 100 ad impressions (one next to each question). Setting this up took a bit more time as I had to select an ad network that was most promising. I settled on Taboola, signed a contract and installed their ads.
Unfortunately that didn’t work either. In a day I generated 30,000 page views (300 people answering 100 questions), but it only generated 3 clicks on the ads! It wasn’t too surprising because people wanted to see what career suited them, not get distracted by irrelevant ads during the process. Moreover, these somewhat intrusive ads impacted my funnel by 10%, meaning that 1 in 10 people now dropped off during the process of answering questions. This was evidence that it was creating a worse user experience.
So I removed the ads and canceled my contract with Taboola, as it was still within their trial period.
Three attempts and three failures. Which was great! I mean, it would have been convenient to make money at this stage. But every test is an opportunity to learn. And I learnt quite a lot from this, especially when I combined it with my previous learning from talking to users at the careers event back in October.
For example, I learnt that users looking for a new career don’t want to spend money on career discovery. Also, while my app is creating value for people, it isn’t enough perceived value on its own.
But my best discovery was how cheaply I was able to acquire users with Facebook ads. This demonstrated that people really wanted to discover the best career for them as they were clicking through the ads, resulting in my getting low acquisition costs. I was also helping people with this problem as they were sharing positive feedback on my ad posts, which resulted in even lower costs. And significantly, this solves a large part of the marketing / distribution challenge I would have in scaling this in an economically viable way. And this last point was also significant in me securing funding …
How I Raised £550k
After a nice long 3-week break over Christmas, I decided my primary goal in the new year was to raise money. This would enable me to focus on this project full-time and maybe even hire a couple of people to join me on the journey.
Nesta’s CareerTech Prize
Back in November, a friend told me about a prize Nesta was running, in partnership with the Department for Education (DfE), for projects focused on building technology to help people navigate their careers.
Nesta is an innovation foundation. They use their expertise and funding to help tackle society’s biggest challenges. As it happens, they were recruiting candidates for their new CareerTech prize. Nesta’s research suggests that more than 6M people in the UK are employed in occupations that are likely to radically change or disappear in the next 10 years due to automation, population aging, urbanisation and the green economy. And coronavirus has intensified the importance of this problem, suggesting that a prolonged lockdown could result in 6.5M people losing their jobs. Of these workers, 80% don’t have a university degree.
So their CareerTech challenge is designed to support people hardest hit by an insecure job market over the coming years. This includes people without a degree, and working in sectors like retail, manufacturing, construction and transport.
They would award 20 companies with £50,000 each to develop their solutions over 9 months, along with lots of input from experts. But the deadline to apply was January 29th!
Applying to the Prize
So I worked full-time on my application over the 10 days I had available before the January deadline. This involved reading up on many research articles, thinking deeply about the judging criteria, looking closely at available labour marketing information and then planning my application.
The application itself was about 4,000 words (14 pages) and I also had to produce a 3-minute video summarising my project. I had a few friends review my application too and provide feedback.
I submitted a day before the deadline and then had to wait 2–3 months for the result.
Raising Money from Investors
Nesta’s prize was a grant, so I didn’t have to give up any ownership in my company. But I might not be successful, and it would help if I could raise more than £50k, especially if I wanted to hire a couple of people to join me.
So I set out to try and raise a further £150k from angel investors (wealthy individuals who invest in software startups). I reached out to a few friends to have initial conversations. The first conversation didn’t go so well as it was felt I was too early and risky (which was fair!). But the second conversation resulted in my friend offering to invest £10,000! That was my first commitment of funds, which was an exciting moment for me!
An Opportunity Through My Founders Network, ICE
I’m part of a network of about 350 tech founders called ICE. It’s incredibly valuable. We have an email list and I’ve asked a dozen questions, and each time I get at least 3–5 responses from experienced founders who had the exact issue I asked about.
One question I asked back in January was what I should know about incorporating a company, as that’s what I was about to do as a requirement for the Nesta prize. As always, I had a bunch of very helpful responses.
One response was from a partner at a VC firm called Forward Partners, who was also a member of ICE. He kindly offered to have a phone call to talk it through. On this call, I shared my company’s vision and what I was doing. He expressed interest and said that I should pop in to discuss further if I was raising money. A perfect opportunity! And serendipitously too!
My First Pitch!
So I had a week to prepare for this meeting, but I had my Nesta application to complete. So the morning of the pitch, I spent 2 hours putting together a pitch deck. I then travelled into London and met with him. The firm’s General Partner (the most senior partner) also joined for the first five minutes, so I was under pressure to deliver a concise pitch to him before he had to leave. I remember visibly perspiring as I had never pitched before and had to do so under time pressure.
Thankfully I managed to say what I wanted to in a reasonably cohesive way and I spent the next hour or so discussing further with the partner there. I even stayed for team drinks as it was a Friday afternoon. All-in-all, I was pleased with how it went, but had no idea whether they would express interest. After the meeting, I quickly followed up by thanking them and sharing my pitch deck.
An Offer to Underwrite £500k
A few days later I had an email inviting me back for further discussions and to meet the other partners. Wow — very exciting! And in the meantime, they asked a few questions over email that took me some time and thought to reply to.
A week later, I returned to their offices and spent a further two hours with them. Half way through this meeting, they said they wanted to invest! Result!
We had some back-and-forth on price, but landed very quickly on what we both felt was fair. They offered to underwrite a £500k round, but there was room for angel investors to invest part of that amount. That was important to me as I wanted good people with strong, relevant expertise to join me on the journey. And I believe the best way to do that is to get them to invest their own money.
So in the end, we managed to secure three angel investors who invested £40k in total and Forward Partners invested the remaining £460k
From talking to other founders, I think I got very lucky. Most founders have to pitch a number of times and sink a lot of time into fundraising. I feel very fortunate to have raised a strong seed round with only one pitch and minimal time spent, which lets me focus more of my time on building the product and company.
Closing the Round
It then took about another 6 weeks to close the round. It turns out that raising VC money requires a lot of legal work. Thankfully, my ICE network came to the rescue again and I was introduced to a fantastic lawyer at a firm called Cooley. It turns out that Cooley represents more than half of the most valuable tech firms in the world. And they were amazing.
Also, the whole Coronavirus pandemic hit halfway through working on the round (more on that later!), which made things a bit trickier. So Cooley really helped during that turbulent time.
And by the end of March, all the money had arrived in my new business bank account!
Hearing Back from Nesta
And literally the same day that I received the money from my investors, I heard back from Nesta. We had been accepted as one of the 20 finalists for the CareerTech challenge! It turns out that they had hundreds of applications, so I feel very fortunate to have made the cut. And that meant a further £50,000.
So in just three months, I went from having no company and no money, to having a company, £550,000, a team of high quality investors to partner with me and access to a whole host of experts in the CareerTech space over the next 9 months.
So I had a pretty epic start to 2020. And then the Coronavirus came …
Impact of the Coronavirus
In mid-March I went on a short ski trip with my ICE network to a resort in the Pyrenees, just over the French border into Spain. In those few short days we were there, the world changed.
Thursday, March 12th
I flew from Gatwick on Thursday and there was very little change in people’s behaviour. It was as crowded as usual with a lot of holiday-makers escaping for a week in the sun. But while we were there, in three short days, quite a lot happened! …
Friday, March 13th
On Friday morning, we heard that the ski resort would close at the end of the day. Spain was closing all ski resorts around the country. So we only had one day of skiing.
Then that afternoon, I had a call from my lawyer strongly urging me to move the fundraising round through to completion as quickly as possible, as the world’s markets were collapsing, which posed a greater risk of not being able to close the round. So I spent a large part of the afternoon reading through the legal documentation and chasing investors and lawyers.
By Friday evening, we heard that the area of Spain we were in had requested the central government of Spain to close all of their roads and borders. If that request was granted, we’d be stuck. So a decision was made for us all to escape over the border to France a day early, on the Saturday afternoon.
Saturday & Sunday, March 14th-15th
We made it over the border and stayed in a hotel in Toulouse, awaiting our return flight on Sunday. Then that evening, in our hotel restaurant at 10pm, the hotel staff informed us that the French government was closing all restaurants at midnight. So we had to vacate by then, and they would provide us with a lunchbox the next day that we could pick up from a table without making any contact.
Thankfully we were able to return to the UK that Sunday afternoon without issue.
The Birth of Coronafriend
At the airport in Toulouse, I was chatting to a couple of my founder friends about the upcoming hardships many will face from being isolated, especially those who are old and vulnerable. We brainstormed how they could be supported, and came upon an idea of trying to build a grassroots movement of neighbourly support.
We reasoned that the best way to support people was their neighbours reaching out to them. And in order to ensure every household in the UK is covered, we could use a map to allow people to “claim” their road once they had posted a note to each of their neighbours to offer their support. The goal was to get every residential road in the UK claimed.
Now it turns out that one of the founders I was talking with had a CTO (Chief Technical Officer) who has a strong background in mapping technology. She looped him into the conversation and on Monday, we all chatted over a video call and made a plan. Thanks to his heroic efforts in building the mapping part of the product, we launched www.CoronaFriend.com just over two weeks later.
This was our attempt to make a difference in this time of great need. Unfortunately it didn’t get the traction we hoped. We had a spike in usage on the first day when we gave it a push, with about 270 roads claimed. This in itself was a reasonable achievement, as that may represent about 15,000 households. But then usage almost disappeared soon after, so we’re a long way from having all residential roads in the UK covered!
But a lot changed in those two weeks! The whole media was taken over by Coronavirus stories, which drowned our story out (the best coverage we got was on Sun’s live feed on their website). We hadn’t built viral growth into the product (such as encouraging people to share with their friends after they claim a road). And as we were all doing this project voluntarily, we all had pressures to return to our day jobs. Plus, schools and nurseries closed in that time, which put further pressure on many of us, me included.
I’m glad we tried to make a difference. And from my experience, maybe 1 in 10 of the products I build gets traction. Plus, we all learnt quite a lot from doing it and I’m hoping a few people got support that otherwise wouldn’t have.
Would You Rather Be’s Focus
And with so many people losing their jobs, the needs of people in the UK have shifted. Plus, there are fewer jobs available and a lot more in certain areas like healthcare. So I’ll probably have to shift my strategy to accommodate these changes and find a way of helping people with the challenges this pandemic is bringing.
Challenges of Having Young Children at Home
The final, big personal challenge I’ve faced with the Coronavirus pandemic is having my two young boys at home with the nurseries and schools closing.
I was a whisker away from going full-time on my startup, but am still having to work part-time. My wife is one of the heroic nurses, so she has to work two long days each week at the hospital. That means I’m needed to care for my boys and attempt to homeschool the eldest. This in itself is very challenging for a whole host of reasons. And that also means I’m only able to work 2–3 days a week on Would You Rather Be.
Thankfully I’m hitting my stride and am fairly productive on those 3 days. So, let me share a little on what I’ve been working on the last couple of months …
My career discovery app seemed to be working well and helping people discover careers that interest them. I’d also had over 10,000 people answer all 100 questions in the app, which yielded over 1M data points that I could analyse and use to make it better. So I figured I’d focus on improving the app — the idea of pouring gasoline on the thing that works.
Also, the company’s mission is to help people discover, get into and flourish in their dream career. So by focusing on improving the current app, I’m making the discover part of my mission better. And that forms the top of my funnel, so any improvements I make here will yield more and happier users when I try to solve the next part of the mission of helping people into their preferred career.
Better Career Coverage
I started by improving the list of careers in the app. Originally, I used a list of 400 careers that I sourced from a single website. I improved this by sourcing a more comprehensive list of careers, starting with the 930 careers the National Career Service have on their site. A few other sources were then used to supplement this.
I then spent some time reviewing this list and consolidating it. Careers that were too niche, would likely lead to low long-term satisfaction or have very low pay were removed. I also combined careers that were similar to each other.
I ended up with about twice the coverage of careers I had in the original list. And I then had to populate my model with data on these new careers.
Improved Career Selection
I spent some time thinking deeply about career selection and whether I should use Machine Learning (AI). I decided that the problem of matching people to careers they would find most satisfying is very complex, as people are complex beings, and each career is incredibly complex in its own right.
When it comes to using Machine Learning, the power of using this depends on the fidelity and quality of the input data. That’s why Machine Learning is really good at image recognition for example, because each image is composed of millions of numbers between 0 and 255, which is a very high fidelity input. But my inputs are just 100 binary answers, which is very low fidelity. So trying to train a Machine Learning model on such a low fidelity set of inputs in order to solve a very complex problem seemed unrealistic.
I knew from user testing that my app is very effective at selecting qualities that match well to someone, but less effective at matching a person to careers. So I decided to continue suggesting top qualities that a person would find satisfying in a career, allow them to pick their favourites and then show them all careers that match their top selected qualities, starting with those that match multiple qualities.
This essentially lets the person choose which careers are right for them from the full list, but helps them narrow down the list of careers based on the qualities they select.
This feature proved very successful and increased the average number of careers that people selected from 4 to 16. It does require more effort from the user, so I also lost 15% more people through my funnel, but I think that’s a worthwhile trade-off to give people far more relevant results.
Improving the Algorithm
The underlying algorithm that powers the career discovery app is proprietary, so I won’t go into detail here as it’s a trade secret. But I built a couple of features that improved the relevance of the algorithm using past data from the 1M data points I already had from 10,000 people completing the app.
One feature was to show a final step of careers not yet shown that other people who are similar to them liked. This feature proved quite successful and resulted in people selecting 2–3 more careers on average.
I also built a lot more intelligence into how pairs are selected, which is the core part of the model. This was using past data gathered, but also new techniques. The improvements here resulted in the number of users asking for their results to be emailed to them increasing by about 35%, which was a good proxy for relevance.
Showing More Information on Selected Careers
Version 1 only showed the list of careers at the end. Version 2 now shows a lot more information next to each career.
The salary range for each career is shown, as well as the ability to sort the final list by salaries. A description for each career is shown as well as its qualities from the ones the user selected.
Each career can also be expanded to show more information. This includes other names this career is known by, other careers similar to it and specific types of a given career. It also shows what each career involves day-to-day.
Ratings & Feedback
Finally, I included the ability for people to share a rating (1–5 stars) and feedback, which has been helpful in informing how helpful people are finding it and for providing input on where to take it next.
About 60% of people have rated the app with 5 stars and 80% with 4 or 5 stars, which is really encouraging. The most commonly requested feature, especially from those with higher ratings, is requesting support into their preferred career. Which is convenient, because that’s the next part of the company’s mission and was where I was planning to take it next.
Since incorporating the company in January and raising money by the end of March, I’ve had a bunch of administrative things to sort out to lay the foundations for the company.
I put together a budget and plan for the next two years. I selected an accountant and got set up with all the tools needed, such as using Xero and Receipt Bank. We also registered for VAT, set up payroll for myself and set up pension contributions. I bought the required insurances (with Digital Risks), registered Twitter, Instagram and LinkedIn handles for the company (in addition to my Facebook page), and set up a new Facebook ad account.
And I set up on GSuite, which is Google’s range of products for businesses. This now means that my company email address is fully operational and all my documents are organised and sorted in my new company’s Google Drive.
Over the past 3 weeks, Nesta has been running a series of workshops for the CareerTech finalists.
We’ve had a session on a study into people who are looking to change careers and don’t have a university degree, so we can better understand their needs and motivations. We’ve had a couple of sessions on business strategy and business development, with a focus on the CareerTech market.
We’ve also had some helpful sessions on service design, which has been particularly interesting and helpful for me. The focus of these sessions is designing services for customers where they get the right information when they need it, are treated fairly and get great value for money. Services that seem effortless for customers, but are underpinned by rigorous research, testing and analysis to ensure they deliver.
I’m guilty of jumping in quickly and creating a product before I do any pre-work. That’s how I’ve started most of my projects over the past two years, including Would You Rather Be. The problem with this approach is that the risk of failure is quite high. Service design is a four-step process, where the fourth step is to create the product. But the first three steps are understanding the context and root problem, imagining solutions and designing, refining and testing concepts. All of this mitigates risk when it comes to creating the solution.
Service Design vs Lean Startup
Now my background and experience is more towards the Lean Startup methodology, which is about creating a minimal viable product (MVP) and then seeing what people do. And there are merits to this approach, as you don’t really know what people will do until you put a real product in their hands that delivers actual value and see how they behave. But there is a cost to building a product, so doing more pre-work up front makes a lot of sense.
I see it on a spectrum, where service (or human-centered) design is on one end, that focuses on a lot of thinking and design work before product creation. And on the other end of the spectrum is the lean/MVP approach of building a product and seeing what people do. Both approaches are trying to achieve the same thing, which is to increase the speed and probability of success of building a valuable product that people use, by learning with as little expense of time and money as possible.
I think where someone lies on that spectrum depends on two things:
- The cost of delivering a solution. If it only takes a small amount of time and money to build a solution, then building it sooner makes sense as you’ll learn more that way. But if it’s expensive to deliver a solution, then de-risking delivery by doing a lot of thinking, research and design up front makes sense.
- The skills on the team. If you have a team of software engineers, building something is straight-forward. But if you have a team of researchers, doing research is more straight-forward and building something might be expensive.
Incorporating Elements of Service Design
I’ve been operating way too far on the lean/MVP part of the spectrum and have never conducted a single formal user interview, for example. So I’ve been influenced by these workshops, as well as input from other advisors and Forward Partners, to spend some time on the research part. So over the next few weeks I’ll be talking to a lot of people and hope to build these skills, which should prove valuable in the future.
And I feel I have plenty of support too! Forward Partners have a studio team that their portfolio companies can use, and they have a lot of experience doing user interviews. So I’ll be working closely with them. And the service design consultancy that has been delivering the Nesta workshops are also providing all finalists with free monthly coaching sessions, which should be very helpful.
So now you’re up-to-date! Over the next few weeks, my top two priorities are user interviews and hiring.
I plan to talk to a range of different people. From users who will be using my products, to employers and recruiters to see if I can figure out a way of making money, to career professionals to learn from them.
I also plan to hire a couple of people. Hiring a senior software engineer / CTO is important so I can build products faster. And hiring a researcher / operations person is important so they can help carry out all the research required to figure out how to get people into their preferred careers and then execute on the operational part of accomplishing this. If you know any good people who would be well suited, please send them my way!
My main goal for this year is to generate income that is scalable and profitable, by solving user needs linked to my mission of helping people discover, get into and flourish in their dream career. That will enable me to be sustainable and then to scale our efforts. And if I can do that with the challenges this pandemic brings, then there should be exciting times ahead!