Speaking at the Big Data & Analytics Summit, Eugene Yan shared his experience leading data science work at Lazada. His central claim is that the field's technical challenges — building models, tuning pipelines, measuring accuracy — are the easy part. The hard work lies elsewhere, in the messy intersection of business goals and human behaviour.
Yan distinguishes between the craft of building a model and the effort of making it matter. On the business side, that means translating vague commercial objectives into concrete problems, negotiating scope, and proving value in terms stakeholders actually care about. These activities rarely show up in textbooks, yet they determine whether a project ships or stalls.
On the people side, the difficulties are hiring the right talent, keeping them engaged, and enabling clear communication between engineers, product managers, and business owners. Yan suggests that teams that master these organisational skills outperform those that merely optimise algorithms. The takeaway for practitioners: invest as much in soft infrastructure as in technical infrastructure.