In a digital world, data is currency. Amos Tay, an experienced Data Analytics Manager, is in the business of coaxing real-world value out of numbers. He hops on the call to let us in on some lessons learned along the way
How it all began
Amos started with a career in accounting and finance, with a serendipitous pivot to data analytics. He joined his first startup, Iflix, and picked up hard skills as a data professional there. With skills in financial analysis, he found himself turning raw data into useful metrics and visualisations influencing high-level decision making. He’s worked at a number of other startups since then, from building the BI team at Fave to setting up the Singapore team of Gojek. These days, he’s running a small team at Rakuten Viki, leading the whole analytics initiative.
NC: When does a tech company start thinking about the data play?
Amos: From day one.
The Prudence of Planning a Data Pipeline
Amos explains that businesses don’t need state of the art data systems from the get-go, but data ought to be on the agenda. He goes on to cite the prudence of having the product well integrated into the data collection pipeline as an example. Companies need different skillsets depending on their stage of growth,
“You don't need to hire a bunch of PhD data scientists the company. What you need are really great analysts who know the data from inside out. Analysts can build basic reports to monitor and they have the ability to deep-dive into what works (and doesn’t) for the business.”
If you’re in the product-market fit stage, you don’t need the flashy tech - just a good firm grasp of the business. The growth of a company’s data systems emerges from these basic tools. At some point, you realise automating easy decisions is sensible. You get people to build those systems. The bigger it gets, the more people you need to maintain the platforms you use.
Amos went through this journey at Fave. They started with Excel reports then progressed to other BI tools to enable people to self-serve data. Governance is the next stage,
“You make sure people understand what information’s available, and that they’re pulling data that suits their needs.”
As a data professional, the job’s partly about teaching everyone how to use the tools but also about evangelising for data itself. At a certain scale, organisations should start embedding analysts into different teams rather than keeping a single centralised data team. Housing skillsets in individual teams helps them pull the necessary data for incident analysis and apply it in a way that’s more productive and effective.
Map → Evangelise → Embed
In 3 words, we sum up what it takes to really weave data-driven culture into the DNA of any business. Map. Evangelise. Embed.
It’s a great framework - to treat data as a horizontal addition, rather than a vertical. But what exactly is a data evangelist? It begins with empathy. There’s really no point in going about preparing a document and telling people how to use Tableau, saying go forth and read. (Spoiler alert: Nobody’s gonna read it). Amos plants seeds - bite-size pieces of information - across open channels for different teams to look at,
“Trigger a little curiosity and they’ll start asking what else they can do with data. That’s the time to get them more involved in the process. Once you have the stakeholders’ buy-in, things shift from a data team being seen as a ‘recipient of requests’ to an indispensable module.
Evangelising a Deeper Data Culture
A key question Amos asks his stakeholders when nudging them through the process is, “What are the problems that you need me to solve for you?”
“What the analytics team delivers is really not numbers on a page. It’s the ability to use data to make decisions.”, he says. Data culture grows deep when there’s constant communication about what you’re doing in analytics and how that shapes each stakeholder’s part of the business.
Interestingly, we’ve got an upcoming course at NewCampus on Stakeholder Buy-in. What does stakeholder buy-in look like for Amos? He begins by talking about goals. A team of highly motivated analysts can cook up a thousand cool projects but most won’t pan out because there’s an alignment issue. What stakeholders want matters. Amos recommends planning a roadmap with their input covering the next 6-12 months factoring in their top three goals or initiatives. Keep a project bank in the back of your head so you can pull up ones relevant to the goals they set. Some projects, like segmentation, fit use cases across different parts of the business - in marketing, operations and finance.
Communicating (is) the value
We turn to the topic of communication and how it’s underappreciated by data professionals as a crucial skill. The industry over-emphasises the technical aspect of data roles. “They build great things, but the value doesn’t translate” Amos concedes. Communication is key to the effective use of information and not just a matter of choosing the right words or format. It’s about an analyst putting themselves in another person’s shoes, seeing their perspective, figuring out what they really want and solving a problem. “That’s a lot more qualitative and subjective. A skill that you do have to build”, he explains. Amos knows this from his own experience from a commercial background.
An ancient method for presenting insight
“What could you have done better? Do you think what you did was understood by the business? If they didn’t, what went wrong?”
Amos’ leadership style involves drawing answers out from his people. These are the questions Amos asks those he leads. Acknowledging his room for growth on the technical side of things, this Socratic approach relieves some pressure and trains individuals to think for themselves. On occasion, he’ll change things up and play devil’s advocate. One trick he’s borrowed from Amazon is to write out narratives. Seeing full sentences on a page helps you see gaps and assumptions in your reasoning when trying to explain something to another person. “It kind of trains the analyst's mind to think in a very narrative format, a sort of storytelling format, something that they can kind of apply to the future”.
Asia’s future data leadership
Data leadership is emergent in Asia and we want to know what Amos thinks the next generation needs to take the industry forward. Amos finds that these days, businesses can get obsessed with the sexy side of data, but most companies don’t need tech like machine learning or AI. He estimates that 80-90% of the business impact is driven by old school, simple SQL queries. Amos suggests that sharpening business acumen and communication skills will make future data leaders invaluable assets.
Final thoughts
We’ll have more sophisticated platforms for data, but understanding the human stories behind the numbers will unlock new possibilities.
- Use the Map, Evangelise, Embed framework to cultivate a strong data culture
- Align data projects with organisational goals and initiatives for maximum impact
- Hone business acumen and communication skills to translate data insights into decisive action
