CASE STUDY · CHILDREN'S HOMES
£23m of residential cost avoidable.
67 fewer children in independent homes.
A county council's independent residential spend was forecast to double in a single year, as more children came into care and weekly costs kept rising. In one month we tested its forecast, benchmarked it against similar councils and built a plan to bring spend back under control.
Objectives and impact
We were commissioned to give leaders a forecast they could trust, show how the council compared with similar councils, and find where cost could be avoided.
OBJECTIVE 01
Produce an independent forecast of residential spend.
IMPACT
2x
this year's budget: independent residential spend could reach £85.5m by March 2028 with no action, tested across four growth scenarios.
OBJECTIVE 02
Show how the council compares with similar councils.
IMPACT
74%
more would be spent on children in care if the council matched its statistical neighbours. The council already spent less than similar councils, so the aim was to keep it that way.
OBJECTIVE 03
Find where cost can be avoided, and set up the council to deliver it.
IMPACT
£23m
avoidable by March 2028, 27% below the forecast, across four workstreams, each reporting monthly to a new sustainability board.
How we did it, and what we found
Child-level placement data, foster carer records and five years of national benchmarks showed where the pressure was really coming from.
Tracked new care entries
+26%
more children entered care in a year, mostly aged 10 to 15, the age group most likely to need residential care.
Reviewed foster carer capacity
47%
of in-house foster places were unavailable, mostly through carer preference. That is more places than the whole independent residential cohort.
Matched carers to children
23%
of in-house foster carers had a child aged 10+ placed with them, even though 79% were open to older children.
Mapped the local market
3x
as many children's home places in the county as children it placed in residential care, with local places about £2k a week cheaper.