# September, in full: 9,500 listing checks, 40 properties analysed, 2 offers, nothing bought
An honest account of one month sourcing in the North East.
—
I want to write down what September actually produced, because the numbers are more useful than the summary. Anyone can say they had a busy month. The question is what came out of the other end.
Here is everything, including the parts that did not work.
The market side: six sweeps, 9,500 listing checks
We pull the entire North East market — every listing across 122 outcodes in the NE, SR, DH, DL and TS postcodes — and run it through the same filter. In September we did that six times.
| Sweep | Live listings pulled | |—|—| | 21 September | 1,864 | | 22 September | 2,101 | | 23 September | 2,097 | | 24 September | 1,637 | | 30 September | 1,627 | | plus the mid-month Property Engine exports | 196 | | Total listing checks | 9,522 |
One thing to be straight about: that is 9,500 checks, not 9,500 properties. We are looking at the same 1,600 to 2,100 live listings each time — the same houses, week after week, with prices moving and new stock arriving at the edges. The market itself is only about two thousand houses in our band. We just look at all of them, repeatedly.
Why repeat it. Because asking prices move, and a house nobody would have bought in August can become a deal in October without the agent ever ringing you. One example from this week: a two-bed in a former colliery village sat at £45,000 for two months, then dropped to £40,000. That £5,000 changed it from marginal to the best yield we have found.
What survived the filter
The band is two and three-bed houses between £40,000 and £110,000, freehold, not auction, not cash-buyers-only, with evidence in the listing that the property is in reasonable condition. On the final full run of the month — 30 September, 1,627 listings — here is where it landed:
| Stage | Count | |—|—| | Live listings pulled | 1,627 | | In band (2–3 bed houses, £40–110k, decent condition) | 273 | | Houses (removed 2 flats and lodges that slipped through) | 271 | | With a rent verified against real local lettings | 269 |
Then the numbers. Three different briefs, because not every investor wants the same thing:
| Brief | Cleared it outright | |—|—| | 8% yield / 15% return / £250 a month | 8 | | 7% / 12% / £200 a month | 10 | | 5.5% / 8% / £100 a month | 70 |
That last line is the one I keep coming back to. Eight houses in the whole market clear a serious investor’s numbers. Seventy clear a reasonable one’s. Same houses, same rents, same condition. The only thing that changed was what the buyer needed — and it changes the answer eightfold.
The part almost nobody does
Passing the yield test is not the same as being a good deal. A house can pay 16% and still be full market value, and if you buy at market value, the yield you bought is the yield you are stuck with.
So for every property that cleared the gate, we pulled the sold prices for its own street — what similar houses on the same road actually sold for, with dates.
Eight houses cleared the yield test. We checked all eight against their own streets. Two survived.
The one that killed most of them:
> A house passed comfortably — 16% gross, £528 a month clear after all costs. Six sales on its street in three years, median £38,200. The asking price was £73,000. It was 91% above the street it sits on.
The yield was real. The value was not there. And the £528 a month was built on a rent that looked like a room-by-room figure rather than a whole house — which is a thing you only notice when you check the street the rent came from.
Two more failed the same way: three-beds asking £60,000 against a street whose median was £50,000. Twenty per cent above. Both cleared the yield gate. Both were expensive.
And one we dropped without valuing it at all: nothing has sold on that street in three years. No comparables means no provable value. I would rather say that than guess, and a lender’s valuer will reach the same conclusion.
The two that survived sat 10% and 21.5% below their own streets. Both are well-evidenced — five sales in three years on one, three on the other. That is what a discount looks like when you can prove it.
The analysis count
Across September, about 40 properties were taken past the first gate and analysed properly — sold comparables pulled and size-adjusted, rents verified against actual lettings, condition judged from photographs, and the full model run at asking price and at an offer price.
Forty sounds low against 9,500. It is the honest ratio, and it is the number that matters. Screening is cheap; proving is expensive, and most of the 9,500 did not deserve the effort. A house that fails the first check does not get the other three.
We also ran two side-scans:
- A low-band scan — everything between £25,000 and £75,000, houses only, 656 properties. This one keeps properties needing modernisation rather than excluding them, because at that price the work is part of the deal. 642 were rent-checked: 90 cleared the 8% brief, 180 the 7% brief, 466 the 5.5% brief.
- An auction scan — 2,104 auction lots catalogued, every September auction in the country, filtered to 31 North East lots. Only four carried an explicit “unsold” tag. We approach auctioneers after the sale for lots that failed, because a vendor who has just watched their house not sell is the most motivated seller there is — but we do not buy at auction, because the buyer’s reservation fee is paid before a surveyor has been in.
The communication side
| | | |—|—| | Emails sent | 664 | | Emails received | 638 | | Distinct agents and firms written to | 118 | | Agent branches asked for tenanted stock | 156 | | Offers put in writing | 2 | | Offers accepted | 1 | | Properties bought | 0 |
Six hundred and sixty-four emails, and two offers. That ratio is the most uncomfortable number in this piece, and it is the one I would change first.
What went out, and what came back
Two offers went in writing.
The first was accepted in principle: a two-bed terrace, £48,000 against a £49,950 asking price, and the vendor said yes subject to a viewing. Then we looked at the photographs properly — carpets worn through, walls needing decoration, a dated bathroom — and the property was not what the buyer’s brief allowed. We withdrew and apologised. The lesson was expensive but cheap in hindsight: we offered before we had a condition verdict, and we should have known better after a fortnight of telling agents we would only send what was ready to let.
The second went in at £48,000 against £56,000 on a tenanted three-bed. The agent never gave us the rent figure, so the offer could not be evidenced, and it lapsed.
Seven properties were put in front of investors, with their comparables, rents and returns attached. One investor engaged properly: he asked for the workings, questioned the tenure, told us he wanted to let to social housing, and then asked for the figures again in more detail. That conversation is the best thing the month produced.
And the thing that explains why no investor has replied until now: we have three newsletter subscribers, and until 30 September we had never sent a single newsletter. Six weeks of sourcing work, thousands of properties analysed, and no way for anyone to hear about it. That was the real bottleneck, and it was not a stock problem at all.
The four things that were actually wrong
1. We screened more than we proved. Nine and a half thousand listing checks produced forty analyses and two offers. The middle of that funnel is too thin — not enough properties carried through to a written offer.
2. We offered before we had evidence. Both offers should have followed a rent confirmation and a condition check. One was withdrawn, one lapsed. Neither was a bad property; both were bad order of operations.
3. The mail filter was eating real agent mail. At the end of the month we found fourteen genuine agent emails sitting in the spam folder, going back eight weeks — two firms had been sending us their property lists throughout and none of it had been read. Some of the conclusion “nobody has stock” was an artefact of our own mailbox, not the market.
4. Nothing was published. The newsletter, the blog and the social posts all existed as drafts and none had gone out. We built a machine and never switched it on.
What changed at the very end of the month
On 30 September we finally shipped: the first newsletter, sent to all three subscribers. A blog post. A Facebook page post with the analysis as a chart. The rent map — every North East district, what houses actually let for, and what that rent can carry — live and free to download.
And the first real investor conversation moved to the point where money was discussed: proof of funds, ID, and a reservation fee. That is the first time in the company’s life that anyone has asked for something of an investor, and it took six weeks of work to earn the right to ask.
The honest summary
September: 9,522 listing checks. 40 properties analysed in full. 2 offers. 1 accepted. 0 completed.
That is not a good month by any measure of output. It is a reasonable month by one measure: we now know, with evidence rather than assumption, that the North East market contains roughly eight houses at a time that clear a serious investor’s numbers, and that most of those still fail on price against their own street. We know the funnel. We know the shape of the problem.
The two numbers I want to move in October are the last two rows of that table — offers and completions. Everything else is machinery, and the machinery works.
—
Kola Bilewu runs Mike Bells Property Sourcing, sourcing turnkey buy-to-lets across the North East. The North East Rent Map — every district, what houses let for, and what that rent can carry — is free at [mikebells.co.uk](https://mikebells.co.uk/the-north-east-rent-map/).