Coliving Pulse #014 — BTR Starts Fell 63% While Completions Rose 21%, and Consent Is Not the Constraint
Published · Hüseyin Şanlıtürk
Contents
Issue fourteen covers 31 October to 6 November. Since issue twelve we have argued that England defines the category first and lets capital follow the rulebook. This week produced a number that sharpens that argument and partly corrects it: the rulebook is working, and the step after it is not.
Savills: 4,696 starts against 21,979 completions
New research by Savills for Real Estate:UK puts UK build-to-rent construction starts at 4,696 homes in the 12 months to September 2026, down from 12,674 a year earlier — a fall of 63 per cent. Outside London the drop was steeper still, at 78 per cent. No absolute regional figures were published alongside those percentages.
Now the number that belongs next to it. Annual completions were up 21 per cent to 21,979 homes, and completions have exceeded starts for the eleventh consecutive quarter. Homes under construction fell 30 per cent nationally over the year to September, including a 28 per cent decline in London.
Put the whole funnel in one line, because the shape is the finding. Consented units increased slightly over the year. Schemes at detailed application fell nine per cent between Q2 and Q3. Homes under construction fell 30 per cent. Starts fell 63 per cent. Completions rose 21 per cent. That is not a sector running out of permission. It is a sector delivering out of a pipeline consented years ago, failing to convert new permission into spades in the ground, and publishing a rising completions number while it happens.
Danny Pinder, director of policy at Real Estate:UK, put the delivery side plainly: "Annual completions are up, which is positive, but if we can't bring new supply into the pipeline quickly enough, we risk exacerbating the UK's housing shortfall, with a significant impact likely to be felt down the line." On the causes, the body's member survey ranked speculation around rent control as the greatest current dampener on investment, with delays at the Building Safety Regulator second; the release notes Gateway 2 delays have improved over the year while delays are increasingly emerging at Gateway 3, "creating a risk that completed buildings remain vacant for longer". Real Estate:UK is calling for targeted tax and regulatory changes, saying the Budget failed to address rising costs.
Two honest caveats, both of which cut against how we are about to use this. First, the research covers build-to-rent and does not mention coliving at all — nor single-family rental or student accommodation. Every inference we draw below from BTR starts to coliving is ours, not Savills'. Second, the reporting we read gives no report title, release date or methodology, so we cannot tell you precisely how "starts" is measured or against which data set. We are publishing the figures because the primary source is named and the internal arithmetic checks out, and flagging the gap because it is the kind of thing that changes a number's meaning.
The same week: three consents, three buildings that lost their use
Southwark Council's planning committee approved an application on the evening of Tuesday 3 November to demolish the Kennington and Walworth Delivery Office at 111-123 Crampton Street and replace it with 146 co-living units. The developer is Bywater, with Waugh Thistleton Architects; the scheme uses a timber frame explicitly to reduce embodied carbon, with communal space at ground level intended to activate the street. Royal Mail announced the sale of the site in December 2025 and intends to leave in 2027. Bywater's own consultation site confirms the approval and describes 146 homes designed to London Plan standards.
Sutton Council approved the redevelopment of St Nicholas House, a former 12-storey office building, into a 359-unit coliving scheme rising to 15 storeys, designed by Assael Architecture. The existing structure will be partially demolished, retained and extended — a retention-led strategy intended to reduce embodied carbon against a full demolition and rebuild. The scheme includes café and co-working space, a permanent home for The Sound Lounge music venue, and around 1,400 sq m of public realm improvements including a new pedestrian route. JRL Group and Halcyon Development Partners are behind it, and also behind a 329-unit coliving scheme on Woking High Street.
The detail in Sutton that we would not have predicted: the application was approved under delegated powers, following more than 100 representations in support and a single objection. No committee, no contest.
Third, and much smaller: changes to a former department store in Dorchester town centre have been agreed, to provide 16 small coliving units with shared kitchens and communal areas plus a gym, at Princes House in Trinity Street. We should say how we read that one — the Dorset Echo's site refuses our requests, so we have it from its own opening paragraph as syndicated across several aggregators rather than from the article itself. The facts above are the ones that lede states; we have left out everything else reported about it.
One more belongs here as a counterweight, because it is an application and not a consent: Sabrecore has submitted plans for a Bournemouth office-to-coliving conversion. Submitted is not granted, and in a week of approvals that distinction is exactly the one that gets lost.
123 Crampton Street public consultation (Bywater) — confirms approval, 146 co-living units
Read →
South London News — Royal Mail Sorting Office Approved for Homes in Kennington (5 November 2026)
Read →
Construction Enquirer — JRL gets green light for 359-unit Sutton co-living revamp (5 November 2026)
Read →
Two headlines, one consent
This week's feed offered "Bywater secures planning for £80m Southwark co-living scheme" and "Royal Mail sorting office in Kennington to be knocked down for 146 co-living homes" as separate items from separate outlets on separate days. They are the same consent. Bywater, Waugh Thistleton, Southwark, timber frame, 146 units, 111-123 Crampton Street. Counted carelessly, the week produced five English coliving consents; counted properly, four, one of which is only an application.
We mention it because trade-press volume is the main way this sector gets described to itself, and the same scheme reported twice reads as momentum. The check that caught it was mechanical: same borough, same product, same architect, same week.
One thing we could not stand behind and have therefore left out: several outlets called Sutton the borough's first coliving scheme. The one article we could read in full does not say so, and Sutton's planning register was unreachable, so we have no council record of the decision, its date or its reference — the approval above rests on a single trade outlet, and we would rather tell you that than imply two.
1.9, not 1.8, and not where everyone looks for it
A secondary source described the Southwark scheme's 146 coliving rooms as counting for roughly 81 conventional homes toward housing supply. That implies a ratio, and the ratio is one of the more consequential numbers an operator can know, so we went after it. We nearly published it wrong in three separate ways, and the corrected version is more useful than the original.
It is not in the London Plan, which is where we looked first and where the figure is usually attributed. It is in MHCLG's Housing Delivery Test measurement rule book, a national document, last updated 17 August 2026. Paragraph 11: "The ratio applied to other communal accommodation is based on the national average number of adults in all households. The current ratio is 1.9." Student accommodation has its own, higher ratio — paragraph 10 gives 2.4, derived by dividing students in student-only households by the number of those households. Both are calculated from England Census 2021 data and will be updated after each Census. The 1.8 figure in circulation is not in the current document, and the 2018 version is no longer available at its old address, so we cannot tell you whether 1.8 was ever right.
Corrected arithmetic, then: 146 ÷ 1.9 is about 77 home-equivalents, not 81.
The third correction matters more than the digit. This is not a planning-policy equivalence that a council applies when deciding your application. It is a measurement device. The test is "a percentage measurement of the number of net homes delivered against the number of homes required", and the adjustment works on the "net increase in bedrooms in other communal accommodation", calculated by MHCLG and added to the net additional dwellings statistic over a rolling three-year period. No officer divides your room count by 1.9 while writing your committee report; the Government divides the borough's bedroom count by 1.9 afterwards, when scoring how the borough did.
Two limits on using it. The document never says "co-living", "shared living" or "large-scale purpose-built shared living" — only "other communal accommodation" — so treating coliving rooms as falling inside that category is an inference, albeit an obvious one. And we have deliberately not told you what follows from a council failing the test, because we did not read that part. If you have been reasoning that your scheme will be waved through because it helps a borough hit its target, the instrument you are reasoning about is further from the decision than it looks.
The thread: the pipeline is rotating, and consent is the easy part
Set the Savills funnel beside the week's consents and our own argument needs adjusting. We have been saying that England's route — define the category, then let capital follow the rulebook — is the one that produces a financeable asset class. The consents this week support the first half of that. Permission is being granted, on brownfield commercial sites, for a product that barely had policy language five years ago. Consented units are up.
The starts number is where it breaks. A 63 per cent fall in BTR starts against a 30 per cent fall in homes under construction says the binding constraint sits between consent and commencement, not before consent. Viability, rent-control speculation, Gateway 3. On that evidence, a coliving consent in 2026 is worth considerably less than a coliving consent looked worth in 2023, and the operator question is not whether a council will say yes but whether anything can be built for the number that made the appraisal work.
The second pattern is about which sites get consent, and it is the actionable one. A Royal Mail delivery office whose occupier announced its own departure. A vacant 12-storey office. A former department store. These are buildings that have already lost their use, and a building with no present constituency has no one to organise against its replacement. Compare Sutton — 100-plus representations in support, one objection, approved by officers without a committee — with Colchester in issue twelve, where a residents' association campaigned against the borough's first coliving scheme on conservation-area grounds. Same product, opposite local politics. The variable is not the merits of coliving. It is whether anybody misses what is there now.
For an operator choosing between sites, that is worth more than most planning advice: the prior use of the building does more for your consent than the quality of your scheme. It also sets a limit on the thesis, because the supply of obviously-dead town-centre commercial stock in good locations is finite, and the cheap consents will go first.
The caveat on all of this: four consents in one week is an observation, not a data set, and the Savills figure is about a different product. We will say so if the next quarter does not fit.
What we are watching
Whether the eleven-quarter run of completions exceeding starts ends, and in which direction. If starts stay down, the completions figure is the next one to fall, and it will fall from a number people are currently quoting as strength.
Whether any of this week's consents convert into a start, and how quickly — particularly Crampton Street, where Royal Mail does not vacate until 2027, so the gap between consent and commencement is built into the programme.
Whether MHCLG ever names coliving in the Housing Delivery Test rule book instead of leaving it inside "other communal accommodation". A product with its own planning policies in a growing number of boroughs, and no name in the document that scores delivery, is a gap somebody will eventually close.
Bournemouth and the other submitted schemes. Applications are the cheap part of the funnel, and this issue's whole argument is about what happens further down it.
A note on this issue
Issue thirteen said in writing that the next issue would be Saturday 7 November. This is that issue, on that day. After changing or missing the stated day three times in six weeks, the only useful thing to report is that the date held, so we are reporting it once and will stop narrating the calendar.

Written by
Hüseyin Şanlıtürk
Founder of StartColiving. Eight-plus years in hospitality and growth marketing, applied to coliving — we build and grow coliving brands, and we built our own marketplace, Rentser.
Published About the author →How we source this →
Sources
- Urban Living News — BTR starts down 63 per cent year-on-year (3 November 2026), by George Sell, reporting research by Savills for Real Estate:UK; figures for the 12 months to September 2026 ↗
- 123 Crampton Street public consultation website (Bywater, with Waugh Thistleton, Gleeds, DP9, Kanda) — confirms Southwark approval of 146 co-living units at 111-123 Crampton Street ↗
- South London News — Royal Mail Sorting Office Approved for Homes in Kennington (5 November 2026); Southwark planning committee decision of 3 November 2026 ↗
- Construction Enquirer — JRL gets green light for 359-unit Sutton co-living revamp (5 November 2026), by Aaron Morby; sole source for the Sutton approval, which we could not corroborate against Sutton Council's planning register ↗
- BTR News — Bywater secures planning for £80m Southwark co-living scheme (5 November 2026); same scheme as the Crampton Street consent above, read only as far as the lede because the article is paywalled ↗
- BTR News — Sabrecore submits plans for Bournemouth office-to-co-living conversion (5 November 2026); application only, not a consent ↗
- Dorset Echo — Former Dorchester department store to get 16 co-living rooms and a gym (3 November 2026), by Trevor Bevins; the Echo's site refuses our requests, so we read only its opening paragraph as syndicated elsewhere ↗
- MHCLG — Housing Delivery Test measurement rule book (first published 24 July 2018, revised 12 December 2024, updated 17 August 2026); paragraph 10 gives the 2.4 student ratio, paragraph 11 the 1.9 other-communal-accommodation ratio, both from England Census 2021 ↗
Growing a coliving brand?
We run growth for 18+ coliving brands — and built our own marketplace.
Book a Free Strategy CallKeep reading
Coliving Pulse #015 — What a Consent Costs, and a City That Won Its Co-Living Case and Backed Down Anyway
Read →
Coliving Pulse #013 — Seoul Has No Legal Category for Coliving, and Morgan Stanley, GIC and CPP Are Already In
Read →
Coliving Pulse #012 — Councils Are Writing Coliving Into Policy While Closing the Door on Converted HMOs
Read →