Bulb with Geraldine de Boisse (Big Bets 6)
This is a newsletter about big bets, explored through conversations with the product leaders who worked on them. For this one, I chatted to Geraldine de Boisse, VP of Innovation at Bulb, the renewable energy start-up. Bulb has 1.7 million members and this year it was named the fastest growing company by revenue in Europe.
Geraldine tells me that Bulb has taken many bets since they were founded in 2013, and offers me three examples to choose from. One was a new line of business, which they managed to launch in a month, one was a great idea that never made it off the ground, and one was a passion product that turned into a big growth driver. Naturally, I want to talk about all three.
It becomes clear over the course of the conversation that there’s no fixed playbook for bets at Bulb, as each example takes a different path, from the source of the idea to the method of delivery. But there are lots of interesting lessons from each.
A bet in a month
Back in 2018, when the company was much smaller, Bulb leadership asked Geraldine and her team to create a new product to pay people who generate renewable energy at home (e.g. via solar panels) for exporting some of that energy to the grid. And they asked for it to be built in a month, so that it could be announced at an upcoming conference.
“We said: ‘That’s not possible’. And Hayden Wood, our CEO said: ‘Try and figure it out’. So we did,” Geraldine laughs, looking surprised - even now - that they managed it. To make it happen, she spun off a cross-functional team of engineers, designers, operations people, and a strategy manager. Their objective was to deliver the absolute minimum functionality required to launch the product.
One month later, Geraldine tells me, they were indeed able to launch a product that calculated the amount of energy someone was able to export to the grid, allowed them to be billed for the energy, and then paid them for it. I point out that this sounds like the whole product, rather than some MVP version of it. ‘A few of the steps were manual,’ she explains. For example, after the sign-up flow, someone from the team would manually get in contact with the new user to onboard them and check the documents. They also leaned on the billing platform from their core business selling energy to customers, so there was little left to do to enable the pay-outs.
As some context on this product, and also the sense of urgency around it: the UK government used to run something called the Feed in Tariff scheme, which paid owners of solar panels for selling energy back to the grid. They were trying to create an incentive for more people to buy and install solar panels on their house. In early 2019, they stopped the scheme, and it was up to private companies to step in and buy the energy if they wanted. That meant there was a ready-made cohort of users for Bulb’s bet.
Did any of them sign up for the newly launched energy export product from Bulb? ‘We had lots of sign-ups after it was announced,’ Geraldine says. ‘But we realised a few were competitors or journalists.’ And today, the service has thousands of customers, a sizeable chunk of the market for energy export in the UK.
We talk about whether the extreme speed was worth it. It certainly gave the project a huge amount of momentum. ‘It definitely wouldn’t have happened if people were working on it with 20% of their time - I don’t think we’d even have the product today,’ she says (alluding to Google’s 20% innovation time). It’s an incredible way to get things done: bringing together a group of people, with all of the required expertise, to work on one specific problem with a deadline.
There were downsides, though. Having shipped the first version in a month, a lot of the next year was spent automating all the manual steps and turning it into a full product. But executives couldn’t understand why the team needed to keep working on it, given that the product already existed. ‘We’d created a gap between expectation and reality,’ she says, ruefully. And it took engineering a lot of time to unpick the hacky version one they’d put into the world.
I ask whether the bet was successful from a business standpoint. Geraldine tells me that it’s only a small part of their business, but points out that the market itself is small, since it relies on people owning latest-generation solar panels. More importantly, Bulb believes that a lot of energy generation will be done at a micro level in future, as well as at big power plants or wind farms. That means these energy exporters are their future customers (‘members’ in Bulb-speak), and the company wants to invest in them now, to learn about them, and be ahead of the curve. ‘It’s an investment for where we think the world will go,’ says Geraldine. ‘After all, electric cars have happened much quicker than anyone anticipated.’ By their own accounting, then, it’s too soon to call this bet.
Carbon calculator
The next bet we talk about started as a hobby project. Most of Bulb’s employees are passionate about reducing carbon emissions, and a lot of them used to calculate their own carbon impact (‘we were geeky and curious’). They got to thinking about whether their members would also like a way to calculate their impact, and Russell, the CMO, realised this could be a great way to bring new members to Bulb. He pitched it to the Senior Leadership team, who were sold after only a brief chat.
Like many companies, Bulb works to a quarterly schedule, but the idea for this bet arrived in the middle of a quarter. They wanted to go after it, but felt they couldn’t deprioritise any of the things the engineering teams were already working on. They did have marketing budget, however, so they decided to put some of that towards an agency to deliver the first version of the product. The agency would do most of the design and research, supported by some of Bulb designers and user researchers, and one of Bulb’s own product teams - the Growth team - would build it.
In some ways this process worked: within 6 weeks, they had shipped a basic proof of concept. ‘But if I were to do it again, I wouldn’t do it like that,’ says Geraldine. In practice, they didn’t really save time by using the agency, because the team at Bulb had to get them up to speed as the subject matter experts, and support their questions. The engineers also didn’t end up delivering some of the other things they’d committed to that quarter, because of the context switching and distraction. And hiring an agency to run a fun new project caused friction with the internal team.
Even if the process wasn’t ideal, the product itself was a success. During the first six weeks of developing it, they tested a number of different concepts with users, including a quick quote version, a version designed for high engagement which had a daily challenge around carbon emissions, and a third version that was more educational. In the end quick quotes won out, but they were able to incorporate elements of each that worked well. At launch, the product just did a very simple calculation of a user’s carbon footprint. They ran social and email campaigns to promote it, and measured its success by how many people signed up and completed the calculation.
But their users told them they didn’t just want to calculate their emissions, they wanted to do something about it. So the following quarter, they added a feature to let people make a monthly payment to offset their personal emissions. After the project with the agency finished, the calculator became fully owned by the growth team, who rapidly iterated on it with a/b tests. For example, they were interested to see if they could improve conversion by removing the personalised element, and just allowing people to pay to offset an amount equivalent to the average carbon emissions generated by a UK resident. It was a success, so they incorporated it into the product as a permanent option, and now 10-20% of users choose to take this route. Offsetting at all was the important thing, rather than trying to be very personalised.
Today the carbon calculator has thousands of paying members, many of whom are customers of Bulb’s energy product, but some of whom are not and just use it as a standalone offering. The product is primarily seen as a way to help the company fulfil its green mission, and - from a business standpoint - improve retention and growth on their core product.
The bet that didn’t happen
As our conversation is wrapping up, Geraldine tells me about a final bet that never went ahead. It started with a solid business case based on saving money. Many of Bulb’s costs are not fully in their control, such as the wholesale cost of energy, so to be competitive in the market you need to keep your other costs down. One major cost is customer service, so Geraldine thought that building their own, highly-automated, customer service tool could be a great way to create competitive advantage. Among other features, a custom tool would be able to route queries correctly to the correct specialist to answer it first time, and surface the relevant member info to any customer service person answering the query.
Geraldine built the case and took it to the senior team, who were interested in the idea, and suggested she speak to a fintech founder who’d pursued the same idea. The fintech founder was extremely negative about their experience, explaining that the project had taken years of work for a large team of engineers, and he wished they’d never started it. He concluded with the advice ‘Don’t do it.’ This effectively killed the idea in the mind of anyone at Bulb. Instead, they took the route of customising existing CS systems, which in the long term has helped it provide excellent service to millions of customers.
‘We’ll never know if it would have paid off, but now it’s much too late,’ Geraldine tells me. In a frustrating coda, a year later she spoke to the fintech founder again who had changed his perspective and decided it was all worth it. This serves as another example of the arc of a big bet - it might feel like a mistake in the middle, before coming good.
Top takeaways:
A tight deadline can sometimes help you achieve extraordinary things, especially if you put a small cross-functional team sprinting on it
Don’t be afraid to fake things behind the scenes for your MVP
Bringing in external teams to work on a bet probably won’t save you time. You might as well prioritise it properly
Be careful how you use others’ advice for your own bets