How do I get my organisation to do data science?

on

So you’re expertly trained in data science, managed to get a great-looking job in your field…. and do basic analytics day to day. Often individuals learn, adapt and innovate quicker than organisations do – so you might find yourself with wicked skills in an organisation that doesn’t know what to do with them. This can be a particular challenge when you’re part of an organisation that has a very small data team (or even just you!). So How do I get my organisation to do better data science? Experience with this myself in several organisations has given me some ideas. Here they are:

  • Explain what you can do
  • Sell your ideas not your skills
  • Create a public profile for your data work
  • Build up a network of helpful, skilled people
  • And collaborate with them

For explanations of each, read on.

Data what?

The first step is sharing the love. You might be fluent in R but, if you’re working in an organisation that doesn’t specialise in data analysis, it’s likely that your colleagues aren’t. They probably don’t know what’s possible. So show them. It’s worth talking to them about what kind of work you can do, how it fits into their agenda, and how it benefits the organisation. Which brings us onto step 2…

WTF?!

A good sell is about more than sharp techniques and new approaches. Spend some time scoping out a few potential projects – and sell them your ideas not your skills. Try to find a single nugget to make your colleagues think WTF?! This might be a killer stat or a counter-intuitive finding that really hits home. You will have to carve out some time to push this, but with luck a little up-front investment on your part will go a long way.

And a bit about attitude. It’s important to carry your idea with confidence because your colleagues will be looking to you to persuade them of the value of these new, foreign ideas. You know the potential benefits of what you are proposing; your audience may not. So don’t be surprised if they are over-cautious or take time to convince. You’re likely asking them to take a risk that they don’t fully understand. Be patient, be persuasive and persevere.

Giselle who?

It helps to build your profile. If others outside your organisation associate you with good data, it’s more likely that the same will begin to happen internally. Build a profile for this type of work and others will come to you for commentary and collaboration. It also shows your organisation that you’re serious, and it’ll help with your next career move. Blog, speak on panels, tweet, work with journalists – whatever gets your work out there.

And I hope this goes without saying but just in case: no bullshit please. Put out work that says something relevant, interesting and new. (And that’s not as hard as it may sound – remember that you’re good at this!)

The good kind of network

I hate networking in the traditional sense and so I don’t do it. What I am suggesting here is finding people with whom you share common ground, and engaging in the genuine and specific sharing of ideas, advice or expertise:

  • Ask experts to peer review your work. They can be subject matter experts and/or those familiar with the analytical techniques you are using. As well as making your work better, this will also get them invested in you and your work. (Remember to acknowledge them!)
  • Ask others for advice. When you come across interesting people in your field, offer to buy them a hot drink in exchange for their thoughts. (A wise colleague once told me: if you want money, ask for advice. If you want advice, ask for money. Try it.) People want to help – and we like talking about ourselves. And a nice byproduct of asking for advice is that you often also get motivation and support too.
  • Crowdsource ideas. What would other people do if they had access to your organisation’s data? What did similar organisations do in other places? Could you apply that interesting approach you saw in that blog to your work? Don’t be afraid of playing with other people’s toys (if you think you need to, ask them first).

Let’s get together

Maybe all those exchanges will lead to something more…. Collaboration is a beautiful thing. Look out for mutually beneficial partnerships (with or without money changing hands). These may be between like-minded individuals, or through organisations like DataKind or Dissertations for good.

Collaboration doesn’t have to happen in the office. If you’ve got the time and inclination, work with organisations that connect analysts / programmers to nonprofits, such as The Operational Research Society and NCRM (researchers/analysts), Social coder (programmers), Cita: IT help for charities and Lasa (general tech support), and Reach and VolunteerMatch (all skills). And of course try offline: Ask around your community – are there any organisations that might want the support of a data expert?

Or try it with money. Sell your skills (on behalf of your organisation) to others who might want to buy in your consultancy. If your organisation won’t release your team for data work – or you don’t have the skills to do it – you could buy it in. Consultancies will be happy to take your cash – but also consider asking relevant academics or other organisations like yours. Show your organisation what good data work can do!

A last word

Try to get to a place where your work programme is best for both your employer and for you. But remember that this stuff can be hard – for you but also for your organisation. It can be a big cultural change and needs time. So keep in mind all the general change management stuff (sounds ridiculous I know but it matters). Keep in mind the importance of ‘hearts and minds’ and try to bring people with you. Don’t forget the bigger picture, and work your goals into those of the organisation. And be patient. After you’ve done all that and nothing changes, and you’re not happy, leave. If you’re reading this post, it’s likely that you’ve got the kind of skills that put you in high demand.

Leave a Reply

Your email address will not be published. Required fields are marked *