The Street That Changed Its Mind

9 September 2026 · Pete Reynolds

In November 1964 a sewer collapsed under London Street in Norwich.

The street had to be shut to traffic for six weeks while it was dug up and repaired. London Street was the city's best shopping street, known locally as the Bond Street of Norwich, and the shopkeepers were alarmed. They expected to lose trade. They worried about deliveries.

Four months later the Eastern Daily Press went back to ask how it had gone. The answer, printed in March 1965: "In the event traders have not suffered and in some cases reported increased trade."

The manager of Jarrolds, the department store, explained why. The shopkeepers had worked out that because cars could never stop on London Street anyway, they had been bringing very little direct trade. Another manager said his customers liked being able to cross the road "with nothing more dangerous than a few holes to think about".

The month after that report, the traders of London Street voted two to one in favour of closing their own street to traffic between nine and six.

That is the whole argument of this article, and it was settled by accident sixty years ago by people who had every reason to believe the opposite.

What Norwich did next

Britain nearly missed it. Through the early 1960s, town after town drew up plans to pedestrianise a shopping street and shelved them. Norwich City Council's own history of the period gives the reason without dressing it up: the plans "had all been shelved after meeting opposition from shopkeepers who believed that their trade would suffer if vehicles were not allowed to circulate through shopping streets."

Norwich had a chief planning officer, Alfred Wood, who saw what the burst sewer had handed him. On 17 July 1967 the Lord Mayor tied a white ribbon across the southern end of London Street and it became the first existing shopping street in Britain closed to traffic. It was a three-month trial. It cost £1,839, kept cheap by making planters out of old kerb sections and sewer rings.

Within a week the cafés had put tables outside. A survey that August counted three times as many people walking down the street as before. Over 90% of those interviewed liked it without traffic and more than 80% wanted it kept.

Then the council asked the shops. Twenty-eight of the thirty that replied said their trade had gone up by between five and twenty per cent. The two that hadn't were national chains reporting a small dip in line with their other branches. The street's six shoe shops all recorded increases of up to 10%, in a year when shoe retailing was falling nationally.

Wood wrote that the experiment "disproves the popular idea of many traders that if vehicles are removed from a shopping street, business will suffer. In fact, the reverse has been found to be the case in London Street."

The scheme was made permanent in 1968 for £25,000. Within three years another twenty British streets had followed. So had Perth, in Australia.

Nobody has proposed reopening London Street to cars since.

The same argument, fifty-six years later

In late September 2020, temporary cycle lanes went in along Kensington High Street in west London, paid for with emergency active travel money and meant to run as an eighteen-month trial.

The Kensington Business Forum and the Kensington and Chelsea Chamber of Trade and Commerce both pressed for removal before Christmas trading. Tom Frost, who chaired the Business Forum, said "the majority of local businesses opposed the current temporary lane". Michael Stone, the Chamber's chairman, said "the cycle lane in its current form is detrimental to business on Kensington High Street".

The council announced removal at the end of November. The lanes were gone by mid-December, seven weeks into an eighteen-month trial.

The reason given was traffic, which the council said had risen 25 to 30%. After the lanes came out, average journey times along the 1.1-mile stretch went up, from 5 minutes 39 seconds to 8 minutes 14 seconds eastbound and from 5:48 to 6:27 westbound. Part of the problem was that cars parked illegally in the space the lane had occupied, as much as 80% of the time.

The cycling numbers had been holding up: 2,384 riders a day in October, 2,897 in November, 2,825 in December, dropping to 2,235 in January once the lanes had gone. Hires from the five nearest Santander Cycles docks were up 14% in October, against a 0.5% fall across the network.

In March 2021 the council voted unanimously not to put the lanes back. A judicial review failed in 2023. Then a consultation that summer drew 1,775 responses, about 66% of them supporting lanes in full or in part, and in July 2023 the council announced they would return after all — as painted advisory lanes, not the protected ones it had removed.

There is an unedifying detail in the middle of this. The Business Forum's original statement read: "We had hoped, like many others that the temporary cycle lanes would have been a success but unfortunately due to the current climate it has not benefited our High Street businesses." The council's press office rewrote it before publication to: "Like many others, we hoped the initiative would be a success. Unfortunately it has not helped our High Street businesses attract customers at a vital time for them, so it is good news that the lanes will be removed." The council said coordinating statements with partners was standard practice.

Why the fear is sincere and still wrong

The most useful study on any of this is small, recent, and German.

Dirk von Schneidemesser and Jody Betzien of the Research Institute for Sustainability in Potsdam took two one-kilometre stretches of Berlin shopping street, Kottbusser Damm in Kreuzberg and Hermannstraße in Neukölln, between June and September 2020. They asked 145 retailers how their customers got there. Then they stood outside and asked 2,019 shoppers.

The retailers estimated that 21.6% of their customers arrived by car. The real figure was 6.6%.

They were wrong about something else too. Asked how many of their customers lived within a kilometre of the shop, they said 12.6%. It was 51.2%. Half their trade was walking distance away and they thought it was one in eight.

What makes this study worth more than the rest is that the researchers went looking for the cause, and found it in their own data. Retailers who drove to work themselves estimated customer car use at 28.6 to 29%. Retailers who got to work some other way estimated 10 to 19%.

The best predictor of what a shop owner believes about their customers is how the shop owner gets to work.

Psychologists call this consensus bias, and everybody does it about everything. It has nothing to do with being bad at business. If you drive in every morning and spend twenty minutes looking for a space, of course you assume your customers are doing the same thing.

It shows up everywhere it's measured

Toronto has the biggest dataset. When the city ran a bike lane down 2.4 km of Bloor Street in 2016, taking out around 160 parking spaces, researchers from the Toronto Centre for Active Transportation and the University of Toronto surveyed 3,005 customers and 525 merchants across three waves, and used a comparison street, Danforth Avenue, that got no bike lane.

Under 10% of customers had come by car. Meanwhile 29% of merchants guessed that a quarter to a half of their customers drove, 18% guessed a half to three-quarters, and 11% guessed more than three-quarters. Between 2015 and 2017 the share of merchants who believed most of their customers drove actually grew by 56%, over the same period that cycling's share of visitors rose from 7% to 20%.

Sustrans surveyed two Bristol shopping streets in 2006. On Gloucester Road in Bishopston, traders overestimated car arrivals by 95%. The real split was 54% on foot, 20% car, 14% bus.

Graz in Austria produced the widest gap anyone has recorded. Actual mode share was 44% on foot, 32% by car, 16% public transport, 8% by bike. Retailers guessed 25% on foot and 58% by car. That figure reaches us through a Sustrans information sheet from 2003 and I could not trace the original Austrian study behind it, so treat it as one step removed, though two separate compilations reproduce it identically.

In Vancouver, Stantec Consulting reported in July 2011 on the separated lanes on Dunsmuir and Hornby Streets. The work was commissioned by the City of Vancouver, the Vancouver Board of Trade and the Downtown Business Improvement Association, none of whom were looking to be told off. Merchants estimated 40% of customers came by car; the real number was about half that. They put public transport at 25%; it was 42%.

New Zealand's transport agency covered nine shopping areas in Auckland, Christchurch and Wellington in 2013, with 1,744 shopper surveys and 144 retailer surveys, and found retailers consistently overestimating how much their trade depended on kerbside parking.

The Danish government's 2023 review of the literature found eight studies that compared what retailers estimated against what customers reported. All eight found retailers undercounting arrivals by bike, by an average of four percentage points, and undercounting arrivals on foot by an average of sixteen.

I could not find a single study anywhere that found retailers underestimating their car trade. The error only runs one way.

Two cases where the shopkeepers were roughly right

They exist, and they matter, because they show what the real rule is.

On Grafton Street in Dublin, traders estimated 13% of shoppers arrived by car. The actual figure was 10%. That is close enough to call correct. On Henry Street, a few hundred yards away, the same 2011 study found traders guessing 19% against an actual 9%. What Dublin's traders got badly wrong was the tram: they thought 28% of Grafton Street shoppers arrived by Luas, when it was 13%, and they undercounted buses and pedestrians heavily.

And before the famous Toronto study there was an earlier one, in Bloor West Village in 2010. The actual car share there was 21%, genuinely higher than in the Annex, and the largest group of merchants, 44% of them, estimated 21 to 50%. Most of them were about right.

So the rule is not that shopkeepers are always wrong. It is that shopkeepers assume their street is a car street, and on most high streets it isn't. Where the street really is car-dependent, they know.

The thing you cannot see from behind the counter

Here is the part that makes this genuinely difficult, and it is the reason the instinct survives so much evidence.

⚡The Weekly Journal · By Email⚡

Urban cycling news, honest stories, and a front-row seat as we rebuild Velorution from scratch. Free, every week.

Join 40,000+ following along

Per visit, drivers usually do spend more. That bit of the shopkeeper's hunch is correct.

Kelly Clifton's team at Portland State University ran intercept surveys at 78 businesses across the Portland area and published in 2013. In bars, people arriving on foot spent the most per visit, about $22, and people arriving by bike spent the least, about $17. But the cyclists came five times a month, which made them the biggest spenders overall at around $82. In restaurants, drivers spent the most per trip, over $19, and made the fewest trips, 2.5 a month.

Toronto found the same shape. After the Bloor Street lane went in, people were visiting three days a month more often than before, while the comparison street saw no change, and the proportion spending $100 or more a month rose from 44% to 53%.

Copenhagen has the cleanest version of the arithmetic. The city's Bicycle Account for 2012 recorded cyclists spending DKK 15.4bn a year in its shops and supermarkets and motorists spending DKK 15.3bn. Effectively level, from two groups with very different habits. Per trip, the cyclists spend much less: a 2018 study in neighbouring Frederiksberg put them at DKK 320 a trip against DKK 620 for drivers. They simply come far more often.

A shop owner sees every transaction. Nobody sees the monthly total per customer. The misjudgement is built into what the job lets you observe.

Don't believe the campaigners' leaflet either

If you run a shop and someone has waved a cycling campaign leaflet at you, some of what was on it was probably rubbish. It's worth knowing which bits, because it is the overselling that makes the whole argument easy to dismiss.

"A protected bike lane raised retail sales 49% on Ninth Avenue in New York." This comes from a real New York City Department of Transportation study published in December 2013, using actual state sales-tax filings, so the number exists. But the 49% is year three specifically; year one was 17%. The Manhattan-wide comparison figure people quote alongside it, 3%, was minus 7% the year before, so the pairing flatters things badly. And the nearby streets that got no bike lane averaged 26% over the same period, which is the comparison that matters. The report also contains two cases nobody quotes: Columbus Avenue, where the bike lane street underperformed the borough, and Willoughby Plaza in Brooklyn, where sales fell 55% in the first year. DOT's own text says the study "does not ultimately prove causality between the street improvement projects and any resulting economic changes."

"Bike parking earns $31 per square metre against $6 for car parking." There is a real peer-reviewed paper behind this, by Alison Lee and Alan March in Australian Planner in 2010. It is about one shopping street in Carlton, inner Melbourne. It gets quoted as though it were a law of cities.

"Sales rose 350% on a street in Seattle." An undergraduate project posted on a blog in 2013, never checked against any actual business. Somebody pointed out in the comments at the time that the spike might be a change in how the state coded its retail data.

"A 1999 San Francisco study found bike lanes helped Valencia Street shops." There is no 1999 study. Valencia Street got its bike lane in 1999. The research is a master's thesis from 2003 based on interviews with 27 merchants.

"People walking and cycling spend 40% more in London's shops than drivers." Transport for London published this in November 2018 and describes it as its own analysis. I assumed when I started reading that it traced back to American research, and that turned out to be wrong too. What is true is that several of the monthly-spend claims printed next to it in the same document come from studies in Portland and San Francisco, and the "up to 30% increase in retail spend" line usually attached to University College London's work actually comes from a different report altogether.

None of this changes the conclusion. It does mean you should be sceptical of anyone leading with a single spectacular percentage.

What the good evidence actually says

Two research teams have gone through everything and counted.

Jamey Volker and Susan Handy at the University of California, Davis reviewed 23 American and Canadian studies in Transport Reviews in 2021. Their conclusion:

"Taken together, the studies indicate that creating or improving active travel facilities generally has positive or non-significant economic impacts on retail and food service businesses abutting or within a short distance of the facilities, though bicycle facilities might have negative economic effects on auto-centric businesses."

That last clause is the honest one. If your business depends on people arriving by car and driving away with something heavy — a carpet showroom, a tyre place, a builders' merchant — a bike lane may cost you money. Follow-up work in San Francisco found exactly that pattern: restaurants and grocers up, appliance and carpet and car-related businesses down.

The second review is Danish, and less comfortable for people on my side of this. EY and Decisio screened 78 studies and kept 52, for the Danish Centre for Cycling Knowledge, in August 2023. Looking specifically at what happened when traffic was calmed or parking removed, they examined 33 separate measures. Fifteen had a clear positive effect on retail trade. Nine had a clear negative effect. Nine had no effect or a mixed one. No category of intervention was reliably good or reliably bad. Local conditions decided it.

Roughly a quarter of these schemes measurably hurt the shops on the street. That is not a number I enjoy typing, but pretending otherwise would be the same overselling I've just complained about.

Two examples of how much it varies. A 2018 study of Barcelona compared two streets that both got cycle tracks, each against a control. On Carrer de Provença the number of shops rose 9%. On Carrer de Londres it fell 16%. Same city, same year.

Ghent rewrote its city-centre traffic circulation in 2017. Passers-by on the two main shopping streets rose from about 480,000 to 510,000, and cycling's share of shopping trips went from roughly 38% to 46%. But walking's share fell from about 32% to 22%, and nearly one in three residents surveyed said they now shopped in the centre less. The evaluators wrote that it was "difficult to conclude from the evaluation whether the mobility plan had a negative or positive impact on the retail trade in the centre of Ghent."

The British numbers

Matthew Carmona's team at University College London studied five improved London streets against five matched ones that hadn't been touched, for Transport for London: Bromley North Village against Orpington, Hornchurch against Upminster, Clapham Old Town against Camberwell Green, Woolwich against Catford, Walworth Road against East Greenwich. They used commercial property data going back up to eleven years.

Retail rents on the improved streets rose 7% a year. On the comparison streets they fell 2% a year. Retail vacancy fell 7% a year where streets had been improved and rose 14.5% a year where they hadn't. House prices barely moved, which is worth knowing if you've heard that this sort of thing is only a gentrification engine.

The same report is blunt about a limit that campaigners skip past. Street improvements on their own had "inconsequential impacts" on traffic or on how people chose to travel, unless road capacity was actually removed. Nicer paving does not get anyone out of a car.

Smaller schemes point the same way. Bognor Regis spent £2.3m on its retail core in 2016 and watched vacancy fall from 13% to 7.5%. Shrewsbury's weekend pedestrianisation produced sales growth 25% higher than the part of town that didn't get it.

One went the other way. On Mill Road in Cambridge, a traders' survey in 2022 found 56.8% of businesses reporting worse performance and 62% reporting fewer customers in the year after a bus gate was installed. That was a bus restriction rather than a cycling scheme, and a survey of opinion rather than a measure of takings, but it belongs here.

The single most rigorous piece of London research is a 2018 paper in PLOS ONE by Klemmer, Brandt and Jarvis, which ran the numbers across 4,835 small London areas between 2013 and 2016 and found more cycling trips associated with more new shops and cafés opening.

The strongest case against me

Vancouver again, because Stantec's 2011 study is the most substantial negative finding in this whole literature and it deserves stating without spin. Merchants on Hornby Street reported a 10% fall in sales and merchants on Dunsmuir reported 4%, which the consultants put at around $2.4m in lost sales and $480,000 in lost profit over a year.

Three things sit next to that. Only four businesses would let the researchers see their actual books, so nearly all of it is what merchants estimated — and where the real figures were checked, the losses were smaller than the estimates. The Hornby lane had only been in for about six months. And retail vacancy on Hornby fell over the same period from 11 or 12% to 2%.

Construction is the other real cost, and it gets waved away too easily by people who aren't paying rent on the street. San Francisco's Controller looked at eleven building projects across the city in 2017 and found five where sales dropped 9 to 19% during the works. If the diggers are outside your door for eighteen months, being told it will be fine afterwards is not much use.

Why I care about this

In 2023 a bus drove into the Velorution shop in Chelsea. A month later a car went into another one.

The business went into administration that August and every shop closed in September. I bought the name, the website and the social channels out of administration that December, which is how I come to be writing this at all.

I'm not going to claim the crashes killed the company. Bike retail was brutal in 2023 and plenty of good shops closed without a vehicle coming through the window. But a chain of shops that existed to get people out of cars was, in its last weeks of trading, hit twice by motor vehicles, and I have thought about that a lot since.

What I'd actually suggest

Not "you're wrong". Something cheaper and more useful: count.

Stand outside your own shop for a week and ask people how they got there. A trader association could do a whole street in an afternoon. It costs nothing but time, and it settles the question locally, which is the only place it can be settled — because the honest answer to "will this scheme help or hurt my shop" is that it depends on the street, and nobody knows your street.

What the evidence does say, without exception, is that the number in your head is probably too high. Every study that has checked has found the same thing, and none has found the opposite.

The shopkeepers of London Street were not fools. They opposed pedestrianisation like everyone else in Britain in 1964. Then a sewer collapsed, they watched what happened on their own street for six weeks, and four months later they voted two to one for the thing they had been against. That is a better response than most of us manage to evidence about anything.