Hilary Mason has been described as a big data rock star. The former chief data scientist for URL-shortening outfit Bitly and data scientist in residence at the venture capital firm Accel Partners now heads up her own consultancy, Fast Forward Labs. Mason is much sought after for her advice in developing new business opportunities through emerging technologies, from supply chain and discovery to marketing.

See also: Pharma adopts data-science culture in move toward AI

Below, she explains her three tips on creating a data-driven culture.

1. Seek leadership support. 

You definitely need it to do this effectively, and if you don’t have it, go somewhere else where you have it.

2. Galvanize the rank-and-file. 

Even if you have leadership support, you also need the support of the people actually doing the work, which is not a given. That’s something that can be encouraged through training, hiring, and exposure.

3. Make the people who resist look good. 

Working with a large banking client on a language-generation project, a client encountered resistance from another internal business lead. A major objection had been, “This technology will never pass with the regulatory group.” When we got the regulatory group to be the first to use it, the client was then able to say, “Look, this is something that’s freeing up lots of people’s time and money.”