When I first trained as a social worker in the early 1970s, I was shown a film entitled A two-year-old goes to hospital, made by James Robertson in 1952. It graphically showed the distress and anxiety of a child in an institutional setting. There were no facilities in those days for parents to stay with their children. This, alongside John Bowlby's 1969 book, Attachment, which developed the work he had carried out during the 1950s on the nature of a child's tie to his or her mother, formed part of our training. Both works demonstrate the profound bond between individuals and the powerful influence of a child's development on the way he is treated by his parents. These theories were at the heart of social work training and practice. For me, safeguarding is as much about safeguarding a child's psychological wellbeing as it is about physical health and safety. So, observing and understanding child and parental behaviour is a core professional activity. Other books have since added to these but, in a nutshell, bureaucratic procedures can be immensely damaging to children. So, why, so many years later and with so much more experience, do inquiries into child deaths frequently expose the lack of direct contact with and observation of the child who is at risk? For me, the role of social workers in the 1970s and early 1980s was much questioned – particularly about the role of psycho-analytic approaches and reliance on the concept of psychological insights as a means of changing behaviour. Sadly, I think this also led to a loss of emphasis on child development, child observation and a proper obsession with attachment as the cornerstones of child wellbeing. My first post as a director of social services was in a council which was still living under the dark shadow of a child death. When I arrived in the authority I was told at least two-thirds of the recommendations from the most recent inspection had been implemented. I was reassured. Two weeks later, a child died in circumstances of chronic neglect that even now I find difficult to comprehend – both the extent of the neglect and more powerfully, how we and other agencies could have failed to spot it. There were lessons to be learned. Lesson number one was that ticking off recommendations to issue guidance and establish systems was not the same as completely changing the way in which professionals behaved in relation to risk. Lesson two was, one had to dig – and keep digging. At least one local office – short of staff and short of leadership – had files stuffed in drawers, cupboards and in trays – all referrals about children potentially at risk but whose cases remained unallocated and unscrutinised – in some cases, dating back 18 months. Putting things right emphasised not only appropriate systems and processes, but also well-trained and well-supported people. The people were key – and always subject to change. Just a few additional vacancies or the loss of a key manager, and the abyss could return quite quickly. I estimate now that three to six months is the times- pan for going from ‘excellent' to ‘inadequate'. That short! Most forms of inspection – no matter how robust – can only give bills of health with a very short shelf life, that the supply of effective professionals needs to be generous and constant, and that vigilance, together with a sense of a constant abyss, are vital components of leadership. As an untrained social worker, fresh out of university, I doubt I was unusual in not realising some families struggled to pay electric bills. In those days, power supplies were frequently cut and vulnerable families were reduced to candle power. I was shocked. I was even more astounded to find one family I worked with allowed their large alsation dog to defecate in the family bath that was also used to bathe their new-born baby. I knew about so-called ‘cycles of deprivation' but the reality of families out of work for generations, the deepseated despair, and the chaotic nature of their lives raised many issues about policy initiatives borne out of rational thinking and benign intent. Modern leaders of children's services need to carry these perspectives, given the numbers of children still growing up in poverty, and they must imbue their staff and services with a strong sense of the reality into which policies and new initiatives are playing. These are just my snapshots. Others will have different ones with different resonances. The point is that our experiences all say something about the mix of perspectives that need now to be brought to bear on leadership and management in children's services in the 21st century. Liz Railton is national programmes director at Serco Education & Children's Services