On Saturday I was in East London for the Not A Gateway festival, a celebration of urban difficulties – a place where artists, urbanists, students and community activists got together to explore the complexities of city life and, importantly, how social justice fits into narratives that are still largely dictated by the property industry.

The festival's hosts, Trenton Oldfield and Deepa Naik, offer a sobering counterweight to the facile arguments of development-led boosterism in their description of the fate of the Cityside regeneration programme. Walk through the redeveloped Spitalfields Market and compare the preponderance of chain stores with the rough and ready local traders on the other side of Commercial Road and it's clear that for many, the constant encroachment of glass-box cities is still what regeneration is about.

I was asked to spend an hour reflecting on what had been learned over the decade and a bit I'd spent at New Start magazine. While there's a book or two to be written about that (and, more importantly, about what we need to do from now on), here's a starter for ten – focusing on what we should have learned but didn't.

1 Regeneration needs to be about building community. Of course, we said we believed that, and in many cases it really happened – from community planning for housing schemes to the creation of networks to support social and voluntary action. But the money went, by and large, into building things: shopping centres and offices and hotels and roads. We need infrastructure, and sometimes on a large scale – but if it doesn't serve the people we have our priorities badly skewed. 

2 Skills and behaviours come first, projects second. Again, there were times when we paid lipservice to this. For a while it seemed people would take the principles of the Egan Review seriously and train up planners and developers in people skills. Instead we resorted again to the blunt instruments, imagining that by creating buildings or businesses the rest would follow. 

3 If you focus on ticking boxes, that's what you learn how to do. We have a government now that says it wants to move away from box-ticking bureaucracy, but that's easier said than done. Organisations, whether in the public, private or charitable spheres, don't tend to reward imagination and initiative because it rocks the boat. Disruptive behaviours aren't encouraged; conformity is. And so we've learned a discourse of innovation and entrepreneurship while squeezing the space required for it to happen. 

4 Entrepreneurship is not about making a quick buck. This is a lesson we've singularly failed to learn. We celebrate wealth as the ultimate indicator of success, and wonder why those who can't gain it legitimately resort to other methods. In the regeneration field there has been much discussion of social return on investment, but few serious attempts to build it into procurement processes. Social enterprise tends to be encouraged not because it puts the social into enterprise, but because it puts the enterprise into social.  

5 The most important assets are the ones we already have. This is true for the built environment, where you bulldoze history and character and memory at your peril; when you do so you need to be confident new history and character and memory can be created with a value that outweighs the old. But the assets of an area are also people and networks and relationships, from the imam to the pub landlord, from the parent at the school gate to the local councillor. More often than not those assets are undervalued. 

6 The people within organisations are more important than the institutions they belong to. We've spent a lot of time learning how to do partnerships – a collection of sworn enemies united in the pursuit of funding, as the cynical saying goes – and we're good at identifying long lists of institutional stakeholders. We are far less adept at identifying the people who can move those institutions – the ones who put their head above the parapet and are ready to take responsibility and take action. 

7 GVA is a terrible indicator of value. 'Gross value added' is a good way of grossing economic activity, but a poor way of identifying what value has been added because it focuses on monetary value alone and assumes not only that all benefits can be quantified in monetary terms, but that what can't be quantified can't be beneficial. And as a result, it reduces the scope for adding the intangible value that is far more fundamental to our sense of wellbeing and purpose in life. 

8 Transformation is a dishonest way of describing short-term change. Political cycles, funding programmes and upward accountability create incentives to over-egg the pudding – to make exaggerated claims about the changes our activities have produced. Among the hackneyed and sometimes mendacious expressions used of run-of-the-mill activity, 'transformation' is the most glaring. Nobody should be allowed to use it until at least 25 years after completing a project. 

9 Not everything that is worth doing is worth doing. This is a more difficult lesson, and it's less surprising we haven't learned it. While much regeneration activity has been shallow and self-serving, there is also much that has been done with the best of intentions and in the most honourable and open of ways, but still shouldn't have happened. The main reason has been over-enthusiasm: we want to do as much good as possible as quickly as possible. But not everything that can grow should be forced. 

10 We never understood we'd need to do it all over again. This is a fundamental lesson of regeneration and one that asks important questions of society as a whole. Regeneration is something we do where a market has failed or an area has declined, and market failure and decline are repeating processes. There is no such thing as a steady state society, and all change exposes vulnerabilities. Regeneration is a way of redressing those vulnerabilities and – if done well – building resilience towards future shocks. The moment we slap a 'job done' sticker on a regeneration project, we sow the seeds of the next failure.