Counted from the tenant's side
Do a landlord's measures show what tenants live through, or only make it look as if everything is fine? The method Stuart Corrigan has worked with since he co-founded Vanguard Scotland in 1999 answers that one service at a time: study the work from the tenant's side, count what actually happens, then redesign it. This page collects nine published housing cases of that method, from 2005 to 2016. The organisations are not named. They describe the same pattern: targets met, tenants waiting, and jobs re-raised or split so that the clock stayed clean.
Each figure is one landlord's own count, as published. None is a sector benchmark.
Repairs
Six published repairs cases, each counted from the tenant's side.
A large Scottish council
Average days per repair, from first contact
Longest repair
No access at the first visit
- What the measures said
- Pre-inspections within three days; job tickets within target.
- How the clock stayed clean
- "Where they cannot get out within 3 days the line is cancelled and a new one issued." Job lines were recorded as complete "although the actual job may not be finished". When a tradesman could not get in, he cancelled the ticket, and the tenant had to "begin at the very start of the process again".
- What tenants lived
- 80% of enquiries were chasing repairs or complaining. 40% of repairs were completed first time. The average took 46 days from first contact, the longest 202.
- Reported after
- Average 4 days, longest 12. No access at the first visit down from 40% to 15%, at the second from 20% to 3.5%.
Vanguard Scotland case study, housing repairs.
A small Scottish council
Routine jobs that came back
Homeless households housed
End-to-end time for repairs
- What the measures said
- "[W]e knew that 95% of routine repairs were completed within 10 days but we did not know that performance varied from 1 day to 39 days."
- What tenants lived
- "[A]lthough the job tickets were completed in target time 24% of routine jobs reoccurred." Tenants typically dealt with 10 different people, and in some cases up to 24.
- Reported after 18 months
- End-to-end time for repairs cut by 66% and for voids by 50%. Repeat repairs down from 24% to 2%. Share of homeless households housed up from 57% to 87%.
Vanguard Scotland case study of a council housing service (2016).
An English housing group
Days for a whole repair
Failure demand
Done in a single visit, after
- What the measures said
- "Despite meeting all targets (both internal and external) for delivery of a responsive maintenance service, 'Check' revealed that the demand for 'I want a repair' was being met with an appalling level of service."
- What tenants lived
- A whole job, "not just a works order", took 125 days. There was a 19% chance that a job would not be finished in the year it was reported. A leak under a bath could need a plumber, a carpenter, a painter and an inspector.
- How the clock stayed clean
- Regulators penalised the landlord for attending when tenants asked, so "staff would cancel jobs and re-raise them closer to the date when the work was being carried out". Under the repair categories it imposed, boarding up a broken window met the target and "tick[ed] the box".
- Reported after the redesign
- Jobs completed within 0 to 12 days, 8.6 on average. 83% done in a single visit, 97% stayed fixed. Failure demand down from 70% to 30%.
Written by the landlord, in Delivering Public Services that Work, Volume One (Triarchy Press, 2010).
A large English council
Days per repair, end to end
Repairs spendAfter: budget for 2010/11
Satisfaction, after
- What the measures said
- An admirable record of accolades and commendations from inspection bodies, while councillors' surgeries were full of tenants complaining about repairs.
- How the clock stayed clean
- "[T]o meet time targets, single jobs were being fragmented into many."
- What tenants lived
- Failure demand was 60% of all demand. A repair took 24 days end to end, and 15% needed four or more visits.
- Reported after
- 7 days on average, 99% satisfaction, over 90% right first time. Repairs spend down from over £35m in 2007/8 to a budgeted figure of under £32m for 2010/11.
Published university case study (around 2011).
A northern English housing association
of repairs contact was failure demand
- What the measures said
- 98% or more of jobs inside target.
- What tenants lived
- The whole repair averaged 28.8 days. 45% of repairs contact was failure demand, and the largest single cause was tenants asking when someone would be coming.
Government evaluation of a systems thinking pilot in housing (2005).
A large Scottish housing association
Lets within four weeks
Rent arrears
- What the measures said
- The repairs contractor's monthly report showed 99% of all jobs completed within the target timescale.
- How the clock stayed clean
- Jobs "that were not completed but nearly at their time limit were being signed off as finished", often as "No Access to property possible", and a "new" job was raised, "so resetting the clock for the target".
- What tenants lived
- 80% of demand from customers was failure demand. Each re-raised job meant a longer wait and another cost.
- Reported after
- Re-let time down from 56 to 43 days, rent arrears down from £10.1m to £7.99m (lettings below). A manager quoted in the case is candid that, beyond performance, "we have not yet gathered real evidence of increased customer satisfaction".
Case study written with the landlord, in Delivering Public Services that Work, Volume One (Triarchy Press, 2010).
Side by side
The same cases, next to each other.
Each bar is one organisation's own count, before its redesign. None is a sector benchmark.
BeforeAfter
Sources as in the cases above.
Voids and lettings
The same kind of target runs empty homes and lettings, and the cases show the same behaviour there.
- Easy voids first. At the small Scottish council, "the void team were focused on hitting their average void target and so when they got behind they would choose houses that they knew required little work to complete to keep their averages down". A void could take 85 days from keys received to the new tenancy.
- Refusals that never needed to happen. The Scottish housing association had over 20,000 refusals in two years, and 44% were caused by the applicant failing to attend or cancelling a viewing. Nearly 1 in 5 new tenancies failed within 12 months. After the redesign, failed viewings fell to 27% of refusal reasons, re-let time from 56 to 43 days, and lets within four weeks rose from 49% to 67%.
- Work started at the deadline. In a government pilot, a contractor at an English arm's-length housing company working to a 28-day target for void repairs would often not start until the target date was near (government evaluation, 2005).
- Re-let in a week. The English housing group, in its own account, reported 80% of empty homes re-let within one week, down from an average of four weeks, and 30% re-let back to back.
Homelessness and adaptations
Redrawn from the published pilot report, co-written by Stuart Corrigan (Vanguard Scotland, 2011). Individual cases are read from the original chart and approximate; averages and limits as published. The pilot cases are shown by their published average and limits, not individually.
A Scottish council: homelessness
People seeking advice assessed as homeless
Homeless households placed in bed and breakfast
- What the system did
- 92% of people who came looking for advice were assessed as homeless. 64% of contacts were failure demand. In a sample of 27 households there were 1,114 demands on the system, and 47% of households assessed as homeless ended up in bed and breakfast. Bed and breakfast spend alone was reaching £1m a year.
- Reported after
- Homelessness assessments down from 92% to 42%, as officers looked for a solution for each household. Placements in bed and breakfast down from around 50% to 15%.
Vanguard Scotland case study, homelessness.
Five English district councils
Average days per adaptation, request to finished workAs in the control chart above
- What the measures said
- Each step was measured on its own, and "No-one was measuring how long it took for the whole process. [...] As a result it looked like the customer was getting a decent service delivered in a timely manner though nothing could have been further from the truth."
- What residents lived
- At one of them an adaptation took 388 days on average from request to finished work, and all that could be predicted was 1,022 days or less. Across five councils the prediction was 936 days or less.
- Reported after
- In the pilot, cases took around 45 to 50 days, and at most 49 once around 45 cases had been through.
Pilot results report, co-written by Stuart Corrigan with a council officer (Vanguard Scotland, 2011).
How to read them
- One landlord each. Every number here is a single organisation's own count. None is a benchmark for the sector.
- Not named. The organisations are described, not named. The published references are available on request.
- Before today's measures. All nine come from 2005 to 2016, before the Tenant Satisfaction Measures in England. Targets of the same kind still run most repairs services.
- Chasing was counted here, not in public data. No public source counts chasing or visits per repair across landlords. These cases are among the few places where it was counted.
"Our satisfaction is high"
The Scottish housing association's own account is honest about this: performance had improved, but it had not yet gathered evidence that satisfaction had risen. High satisfaction says tenants trust the landlord. It does not say the repair cost them little effort. A landlord that does not measure how often tenants chase, how many visits a repair takes and the days from report to fixed cannot tell how much of its satisfaction is patience.
Questions landlords ask
Do repairs targets get gamed?
In four of the published cases on this page, jobs were cancelled, re-raised or split so that each piece stayed inside its target while the tenant's wait went on. Each case is one landlord. The public data cannot show how common it is, because a re-raised job looks like a new one.
How much repairs contact is failure demand?
In the five published cases that counted it, between 45% and 80% of contact was failure demand: chasing, repeat reports and complaints about repairs that were not finished. Each is one landlord's own count, not a sector benchmark.
Are cases from 2005 to 2016 still relevant?
Yes. The measures they describe, such as repairs completed within target and first-time fix, still run most repairs services, and none of them counts how long the tenant waited or how often they had to chase.
