Introduction

Over the last few hundred pages, this book has tried to offer a whistle-stop tour of the What Works agenda – how it came to be, what it has achieved to date, where it has struggled and where it can – and must – go in the future.

We do not intend to rehash the various points made throughout the book in our own words in our summation. Instead, we focus on two things: our tips for starting a new What Works Centre, based on what has been learned both from our own experience and from that of the authors of the book's chapters; and our recommendations for those leading the network to see it thrive into the future.

Top tips for new centres

So, you’ve decided to launch a new What Works Centre – now what?

Carts and horses

Your centre is probably starting in a unique position – so you won't want to copy exactly what one of the other centres has done wholesale. Nonetheless, you need to know what that unique position is, and work out what needs to be done first.

If you’re lucky enough to be leading a financially solvent centre, then you have options about where to begin – with your engagement with policy makers or practitioners – with generating new evidence – with synthesising what is already known. And (especially if you’re a solvent centre), you’ll have no shortage of people telling you which option is most important – usually the one that they specialise in.

Building a relationship with your stakeholders and your sector is important, but isn't a good idea if you have no evidence to show them. Building up an appetite for high-quality evidence, only to let people know that it's half a decade away, is unlikely to win you many fans. Equally, if there's no primary research – like randomised controlled trials and quasi-experimental designs – in your area, commissioning a large number of systematic reviews will only tell you that there's no primary research in your area – which you probably already knew. However, if there are a large number of randomised controlled trials already in your area – as is the case, for example, for parenting programmes – then synthesis followed by rapid communication could be exactly the right thing to do.