Zoopla is one of the UK's leading online real estate platforms, where it has 27,000 partners on the B2B side, 3.4 million listings, 631 million sessions and 21.6 million applicant leads per year. With this much engagement and usage the company is understandably sitting on a mountain of data, which it hopes to make use of in the future to ‘reimagine the real estate industry'.
Antje Bustamante, director of data and analytics at Zoopla, was speaking at the virtual Big Data LDN event this week where she explained that data use at the online platform has historically not been as mature or sophisticated as it could. Antje was brought on board approximately a year ago to build and embed a culture of data across the whole organisation, so that data becomes ‘part of the language'.
Her ambition is to make buying or renting a house as easy as it is to go online and book a hotel or a flight. For anyone that has navigated the complexities of the real estate market in the UK, it's fair to say that this is an ambitious goal. Bustamante said:
We have a purpose to reimagine intelligent home decisions for all. What we mean by that is that everyone of us has probably bought or rented a property in the UK, but from talking about it to people, the process is a bit painful. Zoopla has this amazing chance to change that, to make it better. If you think about everything that's been digitised nowadays, it's incredible that those processes in renting, selling and buying are still so non-digitised. Data is the true enabler to do that.
There is so much going on on our website where people tell us and show us what they want, what they're interested in, what they want to sell, and what sort of people are behind those sessions. Using that data will help us to digitise many of those processes, to guide them along the way. The good news for Zoopla is that we have a lot of data already.
Getting the basics right
Unsurprisingly, Bustamante said that companies Zoopla looks up to for their approach to data include the likes of Amazon, Google and Netflix. However, she noted that businesses too often talk about data in too much of a grandiose way, instead of fixing the fundamentals. She added:
Reimagining any industry is really about having the basics there - that might not be the most fancy thing in the world, but a lot of companies fail exactly there. They have this huge data science team, you hear machine learning all the time, but when you ask them: What are your data principles? Who consumes your data? Who produces your data? There is not a lot of understanding there.
Whereas the likes of Google and Amazon have been successful because they have integrated data and analytics into the very fabric of their business, Bustamante said. It's not just having a data team, everyone in the business is associated with data. She added:
How do they operate? It's not about isolating or thinking I can set up a big data team or hiring 100 people and they will make this company data-driven - every single data process and data culture needs to sit in all of the other teams as well. I think that's a true differentiator. People will say they are data driven, but if you ask them questions like: do you hire people in a data driven way? How do you make data driven decisions? How do you combine qualitative and quantitative data? If we are really honest with each other, not many companies do this in the right way.
When Bustamante joined Zoopla a year ago, she took the time to go and interview 80 of her colleagues to figure out what they want, where the problems lie, and what the opportunities are. Or as she put it, to figure out the "good, the bad and the ugly". The message that Bustamante heard time and time again is that ‘data is hard' - hard to access, hard to draw the right conclusions from, and difficult to ask questions of.
During her maturity assessment, Bustamante considered Zoopla to be ‘data proficient' but not anywhere near ‘data literate' - a rating of 2.75 out of 5.00, if you will. She said:
We are aiming for to be data literate. What that means is to speak data as a language, just like any language. Where everyone in the company can not only use data, but draw the right conclusions from it.
Zoopla is going to achieve this through a number of approaches. Firstly, it is introducing an A/B testing tool to encourage experimentation when it comes to product development. Bustamante wants data at the fingertips of everyone in the company, which is where she believes true transformation will come from. She said:
This is so important because it's about not how we'd change the product from our point of view, but let the consumer tell us. What is the experience they will love? It's about observations, formulating the right hypothesis and continuously improving. Where it also helps, especially if you're using the one tool, is that you have a single framework, you've got shared standards and best practices for everyone in the company.
In addition to this, Zoopla is building a trusted data layer (TDL) in order to bring "power to the people". This TDL has multiple levels to it. At the base is a all of the data sitting on top of a data lake, which is in one place and the data is cleaned, checked and documented. Bustamante said:
What we are doing is combining many different data sources into a raw layer, which can be used by analysts.
On top of that there will be a tidier, more granular layer, which you can access using different tools. However, it is likely that to make use of this layer you will still need coding skills. The next layer up will be a modelled/aggregated/business logic layer, which serves users with all the questions they might need - such as, how many listings are there in Birmingham that don't have a floor plan? Or, how many sessions are there in London at a certain timeframe? Users won't need coding or an analyst for this layer, but a ticket is sent to the data team to get the answers to them.
Finally, right at the top is the self service analytics layer, which will be about providing best in class tooling to enable everyone at Zoopla. Bustamante's team is in the middle of a proof of concept right now to find the right BI tool for this, with the aim of being able to let business users answer 80% of their questions through self-service analytics.
In addition to these approaches, Bustamante is keen to highlight that a successful data culture is about the people. To this end, Zoopla has also made moves to integrate two disparate BI and analytics teams; it is changing its ways of working; it carries out data show and tells; and it has a completely new team structure, where it is embedding analysts within each of the product teams. On that final point, she said:
That's a true, true game changer. We are right in the middle of implementing that, but it ensures that data always has a seat at the company's table. You can't have business impact if you don't know what product is thinking about, what marketing is thinking about, what the problems are of our customers and consumers.
However, Bustamante isn't unrealistic about the challenges that lie ahead. She said that whilst she has an amazing team, as a leader you need to be aware of what road bumps the company faces in the future. She outlined these as follows:
The change in roles and the war for talent - the analyst role is no longer execution, it's such a broad role and in some ways there's no distinction between data scientist and analyst anymore. You need to have more business impact and those people that have that skill set are very rare. You've got to have that on your radar, it's continuously changing.
Need for speed - this is one of the biggest problems at Zoopla. The old data warehouse infrastructures are hard to understand, and Bustamante calls Zoopla's data warehouse "the seven headed beast". You need easily accessible data infrastructure for non-technical people. The success of a company really comes down to this.
From control to collaboration - the understanding of who the data producer is, who the data consumer is. They don't sit within the data team necessarily. You need to talk to the people that produce the data, they need to be aware how to maintain it and clear roles and responsibilities.
The era of the algorithm economy - augmented analytics is one of the things that will come. A lot of jobs, a lot of the things people are doing right now, will be automated, whether we want it or not. You need to understand how fast this is evolving.