Pulling in the Same Direction
A 30 Day Writing Challenge
It’s Day 6 of my 30 Day Writing Challenge, documenting everything I learn as I prepare to start my new job. I’m going to be building a machine learning team, so I’ve spent the past few days looking at how I can mine the data for insights. Today I’ve been thinking more about the actual process of building a team.
I’m joining a funded startup that has been building it’s core product for about a year. It’s well developed and they have a really strong roadmap. My role is to build a team to make sense of the mountains of data they’re generating. There’s no specific goal, it’s simply a case of generate more revenue or cut costs.
I’ve previously written about the economics of software development in Money Going Out and Money Coming In. However, this is a much bigger task. The team is completely new. I’ll have to justify its existence, its budget, its size.
Everything I do will be experimental. “Can we affect x by doing y?” It’s really exciting, but it’s crucial that I set myself up to succeed. Any fast-growing startup, backed by venture capital, lives and breathes by its metrics. It’s important that I choose the right metrics to measure.
There are things you can measure which are impressive numbers, e.g. number of app downloads. What does that number really mean though? How can you use that number to improve the product? We call these vanity metrics. They help inflate your ego but aren’t useful for steering the product.
In contrast, actionable metrics are numbers which really illustrates how each release makes your product better. It might be average revenue per transaction or cost per conversion. It’s important that you can use your data to make decisions, not just to feel good.
We’ll be running lots of experiments using various machine learning and data mining techniques. It’s critical that we build features with a hypothesis about which metrics will be affected. For example, when we implement the strategy discussed in Finding Cash Cow Customers in Your Data we might expect to see an uptick in average customer lifetime value. Once we push a feature out I’ll be monitoring the metric to ensure our hypothesis was valid.
But it’s not just for me to monitor the metrics. The team should be empowered to work on whatever will add most value to the business. It makes sense to put the metrics front and centre. One of the first tasks will be to create a dashboard which displays charts of the key performance indicators over time. The charts will be annotated with each feature release so it’s clear, at a glance, how each deployment affected the metrics. If the entire team can see the numbers, they’ll have a good sense of how our work is affecting the business.
Only by measuring the key performance indicators from day one and designing product features around the metric they’ll affect will I set myself up to succeed. I want to demonstrate added value from the minute I arrive, and at every subsequent opportunity. If I can do that, I’ll successfully build a new team that is as valuable to the business as the core team.
This is a post in my 30 Day Writing Challenge. I’m a software engineer, trying to understand machine learning. I haven’t got a PhD, so I’ll be explaining things with simple language and lots of examples.
Follow me on Twitter to see my latest posts. If you liked this article please click the heart button below to share — it will help other people see it.
