Head of BI

Nous is an AI agent that takes some of the load of life; making good decisions and acting for our users in areas where they’re not the experts or aren’t paying enough attention. We’ve started with optimising bills, where we’re already saving households thousands, and are now expanding into new categories. Similar to how Amazon picked books as its first category and grew from there.

We’re growing fast, and it’s working. 20x YoY growth makes us one of the fastest scaling startups in the UK. And NPS of +70 (higher than Apple) shows that our users love it.

We’ve just launched two more novel new products (on top of our existing bill optimisation). Insurance: we’re the first to get the regulatory approval and build the tech to actually do the full insurance discovery and purchase journey as an agent for the user (not just give them a link to the best deal). Subscription management: lots of companies use Open Banking to draw pretty graphs, we’re the first to use it to actively save people money starting by cancelling unwanted subscriptions for them.

It would be easy to look at this and think we’ve got it all figured out. But we’ve only scratched the surface; our ambitions are SO much higher.

We’re building a new category of product which is hard, so we’ve stacked the deck in our favour: an experienced founding team with multiple exits, and investors including the founders of Monzo, Wise, Booking.com, lastminute.com, Onfido, Funding Circle, Tide, Habito and more.

Still, the problems we’re solving are hard. ‘Talk is cheap’ applies to agents too. To have REAL impact, agents need to break out of the sandbox and take action in the real world where there are consequences. This means solving all the hard, thorny, interesting, frustrating problems the real world has to offer. ABOUT THE ROLE

Across energy, mobile, broadband, insurance and subscriptions, Nous has grown 20x in the last 12 months. Our reporting hasn’t always kept pace. We’re hiring a Head of Business Intelligence to own this and fix it.

That means you’ll set the standards rather than inherit them. No legacy team, no committee, no five-year-old semantic layer you’re not allowed to touch. Which cuts both ways.

Take our unit economics. Revenue per household is a commercial assumption rather than a column in a table, and it varies as a function of behaviour, service lines managed, engagement. Some users are Premium, some not; some suppliers pay a bounty per switch, some pay nothing. Households change providers, cancel services, then switch again, and their value compounds over time in a complex way. Getting from a decent estimate to a genuinely defensible view of LTV, payback and cohort behaviour is one of the highest-leverage things anyone could do here, and it’s the kind of problem that has no textbook answer.

To get there you’ll own the data foundations too: canonical definitions of our fundamental metrics, the dbt models underneath them, and the reporting that Growth, Commercial, Operations and Finance depend on. We’ve built quickly and there’s real judgment and craft required to make that as solid as the business now requires but doing so in a scrappy and pragmatic way.

You’d work directly with the senior team: CEO, Head of Finance, and the Growth, Commercial, Operations and Product leads. Your work is what the board sees. Our current stack is Postgres replicated into Snowflake via Fivetran, dbt for transformation, Omni for modelling and dashboards, Amplitude for product analytics, and Statsig for experimentation, plus a growing set of bespoke dashboarding and reporting instrumentation., * Own the canonical definitions. Sponsor an agreed definition for each fundamental metric. Then hold the line intelligently when someone wants a subtly different version for their deck.

• Own the models underneath. Our dbt and Omni models have grown alongside the business, and there’s a satisfying piece of work in consolidating them, moving everything onto our current event standard, and retiring what’s no longer earning its place.
• Make the numbers provable. Events, database tables and the finance ledger should agree, and you’d build the reconciliation and automated checks that keep them agreeing as we scale.
• Make self-serve work properly. A clean semantic layer, sensible naming, documentation people actually read, and a small set of dashboards trusted enough that nobody builds their own version on the side.
• Own commercial reporting and unit economics: revenue per switch by product and supplier, contribution margin, CAC and payback by channel, cohort retention, repeat switching, LTV. Partner with Finance on management reporting, the board pack, and the reporting infrastructure behind investor updates.
• Be the analytical partner to the commercial teams. Growth needs to know which channels work. Commercial needs to know which supplier deals are worth doing. Operations needs to know where the process leaks. Answer those questions, then build the reporting that stops them being asked again.
• Set the governance: access, PII handling, how deleted users are treated in analytics, what can and can’t leave an aggregate. Plus the commercial side of the data stack itself, since pipeline spend at our scale is worth someone owning properly.
• Continue to drive the use of AI. Most of our analytical work now happens in Claude Code alongside SQL. We’d expect you to continue to push this hard..

ABOUT YOU

You’re the person who noticed the number was wrong. Somewhere you’ve worked, there was a metric everybody quoted that didn’t survive contact with the underlying data, and you were the one who found it, worked out why, and fixed it properly instead of adding a footnote.

You’ll likely thrive in this role if:

• You’re commercially minded as well as technically strong. You can build the model, and you have a view on what it says. When you present a cohort curve you have an opinion about what we should do about it.
• You’re fussy about correctness in a way other people find slightly excessive. You triangulate. You reconcile. You won’t ship a chart you can’t defend, and you’re comfortable pushing back on a number senior people have grown attached to, because you’ve already checked it three ways.
• You’ve done this from scratch. You’ve walked into a fast-growing company where the reporting was held together with string and made it solid. You know what order to do it in, and which bits can wait.
• You’re hands-on and expect to stay that way. You’ll be writing the SQL and building the models yourself for a good while. We keep the middle thin on purpose.
• You simplify. Faced with forty dashboards, your instinct is to work out which six matter and kill the rest. You’d rather have a few things everyone trusts than a complete set nobody opens.
• You build with AI tools daily. Not “I’ve tried ChatGPT”. You use Claude Code or similar as part of how you actually work: writing and debugging SQL, navigating an unfamiliar schema, automating your own checks.

• Substantial experience in analytics, BI or data leadership, including a spell owning the numbers for a fast-growing consumer or transactional business. We care more about what you’ve built than the years attached to it.
• Confident SQL. Handed an unfamiliar schema, you can get to a defensible answer without help.
• Hands-on with a modern warehouse (Snowflake, BigQuery, Redshift or similar) and a transformation layer (dbt or equivalent). You’ve written and maintained the models, not just commissioned them.
• Hands-on with a BI and semantic modelling tool (Omni, Looker, Lightdash, Metabase, Power BI, Tableau), including designing the semantic layer rather than only building charts on top of someone else’s.
• Event analytics (Amplitude, Mixpanel, GA4 or similar), and a working understanding of where event data and database data disagree and why.
• A track record of establishing metric definitions and getting an organisation to actually adopt them. That’s as much a persuasion problem as a technical one.
• Comfort working directly with finance on management reporting and reconciliation.
• Active use of AI tooling in your day-to-day work., * Consumer subscription, marketplace, fintech, energy or utilities. Anywhere with messy third-party data and revenue that isn’t a clean row in a table.
• Modelling LTV, payback and cohort behaviour where the inputs are commercial assumptions rather than observed values, and being clear-eyed about the uncertainty that creates.
• Experimentation platforms (Statsig, Optimizely, GrowthBook) and reading experiment results honestly.
• Board and investor reporting.
• Instrumentation and tracking plan design, including working with engineers to get events fired in the right places and keep them there.
• Some Python for the analysis that doesn’t fit in SQL.
• You’ve built your own tools when the ones available weren’t good enough.

REMUNERATION AND BENEFITS

• Compelling packages, including share options with life-changing upside. We’ve raised enough money to pay real salaries. We’re building something great together and we all share in the upside. If we win, we win together.

• An environment where you can do your best work:

• A great office set up for lots of collaborative working. We already need more whiteboards
• The equipment / setup / training you need to do your best work
• A high-context environment with lots of exposure to the big decisions - be in the room when it happens (literally)
• Flexibility around remote work. If you sometimes just need to get your head down, that’s fine
• Very good coffee
• ️ A well-stocked kitchen (featuring a wildly popular toastie maker) so you can always make yourself a breakfast or light lunch on the house
• Informal social every week, something bigger every month, plus a set-piece event in the summer and again at Christmas - we’re not robots!
• 33 days paid holidays (including bank holidays)
• Health and dental cash plans, including an employee assistance programme
• Travel benefits, including the cycle-to-work scheme and season ticket loans

Nous has an office-first culture. That means we encourage in-person working, and most people do work from the office in Farringdon most of the time. We do a lot of brainstorming, problem-solving and interaction within the team and presence makes this more effective. It also helps ensure people have good peripheral awareness and high context. But we are pragmatic not doctrinal about this, and recognise that, sometimes, certain types of work can be better done remotely. THE ENVIRONMENT

Most importantly, this is a high-performance, high-agency and high-trust technical environment. Nous is intentionally lean and AI-native, not just in its product but as a company. Everyone is expected to use agentic tooling to its full potential, whether that’s debugging a system, streamlining a workflow or improving how customers are served. Rather than building layers of middle management, we hire smart, high-agency people and give them real ownership.