Impact

Align the efforts of your team with the growth engine of your company.
Data is a standalone function at the same level as engineering, product or marketing. Not a sub function, not a ticket answering unit.
Data needs a seat at the decision table if you want to operate your business smartly.
Cultivate a culture around evidence and structured decision making. It will save you time later on to have allies understanding the importance of data, well formed hypotheses, a good testing protocol and the difference between data science and wishful thinking / confirming gut feelings.
Spend time making good architecture and infrastructure decisions. Interoperability, anti lock-in, maintainability, how long it takes to operate it, should be factored in, in addition of price.

Productivity

Say no to most things thrown at you.
You only got a few opportunities each year to deliver massive value / swing for the fences. The rest of the time is routine work to support others. Make sure you protect the time when you’re building.
Don’t fall prey of “busy-ness”. Your output doesn’t follow a normal distribution where each task adds a bit of value. It’s a power law.
Your company doesn’t need more dashboards. You’re better off having fewer of them but people looking at them.
Never lose track of the business value when working on infrastructure projects.

Be in a company where data is a profit center.
The data ecosystem is still young. Many things suck and most tools hardly work together.
Be in touch with your peers from other companies.
Some companies are profitable enough and/or poorly managed enough so that data can become a bullshit job. Spot them and avoid them.