Brief / Checklist

The ground plane checklist

Forty-five checks, sorted by one question: can you still undo it?

The drawings are on the wall. The awning line is being discussed. Someone asks whether the paving should run through the crossing or stop at the kerb, and there is a productive twenty minutes on that question. The meeting ends. Nobody has asked what a tenant on that frontage would have to turn over to pay the rent in the feasibility.

This is the ordinary failure. Not bad design. Good design, reviewed against the wrong questions.

Every design review we have sat in spends most of its time on the things that are easiest to have an opinion about. Bench placement. Paving colour. The signage family. Everyone in the room can contribute to those, so those are what get the hour. Almost nobody challenges a footpath dimension at concept stage, and the footpath dimension is the one that cannot be undone.

So this checklist is not sorted by topic. It is sorted by what you can still change.

Twelve you cannot undo

Set in the built form. Get them wrong and no amount of leasing, programming or management recovers it. These are worth the whole hour.

Nineteen the deal locks

The mix, the leases, the rent model, the governance. Changeable, but only at renewal or renegotiation, which means five to seven years.

Fourteen you fix on a Tuesday

They matter. They are also cheap and fast to change, which makes design review the wrong place to spend time on them.

How to use it

Take it into a design review, not a strategy workshop. If you only have twenty minutes, spend them on the first twelve.

Mark each item one of three ways. Resolved means there is a written answer somewhere you could point to. Assumed means you think you know but nobody has tested it. Not applicable is a legitimate answer, used deliberately.

The output that matters is the list of things you marked assumed. That is your assumption register: the things this scheme is currently betting on that nobody has checked. Every one of them should get a test and a date before capital commits.

Your marks stay in this browser. Nothing is sent to us unless you ask for it by email.

What this is for. Mixed-use, retail-led ground planes: town centres, high streets, build-to-rent, mixed-use precincts and transit-oriented development. Health, education and workplace precincts run on the same methodology but need a different subset and their own anchor questions. Ask us for those.

Show

Tier one: the twelve you cannot undo

Locked by the built form. The only remedy after construction is demolition.

T1-01

1Has every street, lane and through-link been measured and assigned a band?

Benchmark

Our Street-Width Monogram runs A at 1.75 to 2.5 metres for micro-lanes, B at 2.6 to 3.5 for intimate lanes, C at 4.0 to 5.5 for fine-grain clusters, D at 6.0 to 7.0 for shared-way spines, E at 9 to 11 for pocket streets, and F at 12 to 18 for urban main streets.

Fails when

Widths are set by the traffic engineer and the retail is asked to fit. Width is destiny. A 9-metre lane will not behave like a 3-metre one no matter what is leased into it.

T1-02

2Is the street enclosed, and is it two-sided?

Benchmark

Width-to-height ratio of 1:1 or greater. Successful high streets are typically around 20 metres wide or less.

Fails when

The street is wide, the buildings are low, and one side is a residential lobby and a substation. Single-sided retail halves the reason to walk the length, and low enclosure makes the space read as a car park with shops around it.

T1-03

3Are footpaths wide enough for two-way flow plus alfresco, at the built pinch point?

Benchmark

3.5 to 4.5 metres in active frontage.

Fails when

The drawing shows 4 metres and the built pinch point is 2.2 metres behind a planter and a bin. Draw the section with a pram, a table and two people passing in it.

T1-04

4Is there a shopfront rhythm, and is the glazing under control?

Benchmark

A door or a change of frontage every 4 to 8 metres, with facade variation every 3 to 8 metres. Glazing below roughly 60 per cent, with layered display behind it.

Fails when

One tenant holds 30 metres of frontage behind full-height uniform glass. The walk past becomes a walk through, and the operator has nowhere to merchandise.

T1-05

5Do tenancy proportions work for the operators named in the mix?

Benchmark

Roughly 1:3 shopfront-to-depth. Boutique 15 to 45 sqm, anchors 50 to 100 sqm, micro tenancies 20 to 50 sqm as incubators, pop-up spaces around 25 sqm.

Fails when

Deep narrow tenancies are created because they suited the structural grid. They will trade poorly and they will discount at renewal, permanently.

T1-06

6Are service, loading, waste and back-of-house resolved without consuming active frontage or a corner?

Fails when

The loading dock is discovered late and takes the best corner. This is the most common late-stage frontage loss we see. A corner sees two streets and should carry a tenant that benefits from both.

T1-07

7Is there continuous canopy, shade and rain shelter along the primary walking route?

Benchmark

Street trees are associated with roughly 20 per cent higher retail income on treed streets (Burden), around 40 per cent fewer crashes in one Texas study, and 40 to 60 per cent longer pavement life (UC Davis).

Fails when

Trees are placed for the landscape drawing rather than the walking route, and the awning is discontinuous. Weather protection on the main route determines winter trade more than most design decisions.

T1-08

8Have colonnades been rejected, or justified?

Fails when

A colonnade is drawn for weather protection. In Australia they usually darken shopfronts, break sightlines and create surveillance problems. The exceptions need height, very high foot traffic and immediate street adjacency. Default to no and make the case if you want one.

T1-09

9Do entries connect to the routes people already walk, and can a first-time visitor read the place from them?

Fails when

The precinct is internally coherent and turns its back on the desire lines around it, and wayfinding is expected to compensate. Sightlines to a visible destination do more than signage ever will. Watch where people currently walk and put the entries there.

T1-10

10Has the main public space been located and oriented for sun, wind and sightlines, rather than for what was left over?

Benchmark

The comfortable range is roughly 21 to 27 degrees. Test winter sun access and summer shade on the spaces people are meant to sit in, and test wind at the alfresco locations.

Fails when

The main space is in permanent shadow from 2pm, in full western sun in February, or under a podium downdraught. Any of the three empties it, and none is fixable afterwards.

T1-11

11Has closing the street to traffic been properly interrogated?

Benchmark

Judge (2015) reports that only 11 per cent of installed US pedestrian streets are deemed successful, and NACTO indicates roughly 12 per cent of the original 200 remained fully operational by 2005. Australia built 48 between the 1970s and 1980s and at least 10 have reopened to traffic. Auckland Council's 2012 evaluation of Fort Street, run as a shared street, recorded a 54 per cent increase in pedestrian volumes and a 47 per cent increase in consumer spending, with zero crashes in the post-implementation period and 42 per cent of pedestrians reporting they felt safer than before. Horseshoe Common in Bournemouth recorded footfall up 126 per cent and casualty collisions down 38 per cent (Pearson, Philp and Whyte, ICE Municipal Engineer, 2019).

Fails when

A street is closed because closure feels like the pedestrian-friendly choice, or speed is managed with a sign on a pole while the geometry says 50. Where pedestrian volumes alone cannot sustain trade across the day, a shared street outperforms a pedestrian-only pathway. Our Band D shared-way spines are designed to a 10 km/h maximum, and the geometry has to deliver it.

T1-12

12Is the publicly accessible space dedicated and designed as an asset, rather than surrendered as a concession?

Benchmark

Parks and open space lift nearby residential values by an average of 2.6 per cent, rising to 6 to 7 per cent for high-impact spaces with active programming and good design (Alluvium and City of Moreland, 2020). A Brisbane hedonic analysis projected a 3 per cent price increase within 750 metres of converted parkland, concentrated in the first 400 metres (ScienceDirect, 2022).

Fails when

The public space is the floor-space bonus made physical, sized to the planning requirement and located where it cost the least. Well-designed publicly accessible space is among the highest-returning investments a precinct developer can make, and the returns are set by where it is and what faces it, both of which are locked at this stage.

Tier two: the nineteen the deal locks

Set by the mix, the leases, the rent model and the governance. Changeable at renewal or renegotiation. Assume five to seven years.

T2-01

13Is there a strategic position on one page that names what this ground plane is not for, and does the plan hierarchy enforce it?

Benchmark

Readable in 90 seconds, understood by an investment committee in three minutes. The hierarchy runs development strategy, then ground plane strategy, then the retail, public domain, activation, servicing, leasing and marketing plans.

Fails when

The positioning welcomes every use, or leasing targets are set before the ground plane strategy exists. Leasing without strategy is building walls without a foundation.

T2-02

14Is it explicit which of the three user horizons drives the design?

Fails when

The project claims daily users, visitors and destination seekers equally. Daily users are the economic backbone and want convenience and reliability. Destination seekers want something worth the trip. Those two briefs pull design in opposite directions and one of them has to win.

T2-03

15Is there a written list of the conditions under which this strategy fails?

Fails when

Assumptions live in people's heads. Every load-bearing assumption should sit on one sheet with its evidence, the test that would break it, and a current verdict. If you are marking this checklist honestly, you are already building the first column of it.

T2-04

16Is the catchment read qualitatively as well as quantitatively?

Fails when

A floorspace gap analysis says there is no gap and nobody has walked the incumbents. A theoretical absence of gap routinely conceals poor operators in poor locations. That is a competitive gap rather than an arithmetic one, and it is a finding.

T2-05

17Has anyone established why a person would come here rather than the place they currently go?

Fails when

The answer is "there will be 4,000 new residents". Every precinct competes with the couch, the local strip and the phone. Residents are demand, not a reason.

T2-06

18Which of the six layers of experience are included, which are absent, and why?

Fails when

Nobody has stated the absences. A strategy that lists what is present and stays silent on what is missing has not made a choice, it has made an inventory.

T2-07

19Has the everyday needs provision, including the non-retail layer, been tested against catchment population?

Benchmark

Our Everyday Needs Benchmarks run roughly one supermarket per 8,000 to 12,000 people, one pharmacy per 3,000 to 5,000, one cafe per 1,000 to 3,000 with a strong clustering effect, one GP per 1,200 to 1,500, one gym per 5,000 to 10,000, and one childcare place per five to seven children aged nought to five.

Fails when

The mix is copied from a comparable scheme with a different catchment, or medical, allied health, childcare and services are treated as filler for the tenancies retail did not want. They generate the most reliable weekday traffic in most precincts.

T2-08

20Does the food and beverage offer reach critical mass, in one place?

Benchmark

Five to eight venues create a cluster people will travel for.

Fails when

Three restaurants are spread across three separate frontages so none of them benefits from the others.

T2-09

21Is there a minimum viable opening position?

Benchmark

At least six food and beverage tenancies trading and 70 per cent occupancy at opening.

Fails when

The precinct opens 40 per cent leased and spends three years fighting the impression that it is empty. This weighs heaviest in build-to-rent, where the community does not exist on day one.

T2-10

22Is rent built from what operators can trade, and has the mix been tested with people who would actually sign?

Benchmark

Establish the turnover productivity range for each category, apply a category-appropriate occupancy cost ratio, and derive the achievable rent. For food and beverage under turnover-linked structures the working benchmark is around 7 per cent.

Fails when

Rent is set top-down and the categories have never been shown to a real operator. A 12.5 per cent rent premium means nothing if the tenant cannot achieve the sales to support it. This is the single most common cause of a well-designed ground plane failing.

T2-11

23Is the anchor identified, and is it the right kind of anchor?

Fails when

The anchor is assumed to be a supermarket. Libraries, cultural institutions, transit interchanges, education and health facilities and food halls all now function as anchors, and several of them generate more useful dwell than a supermarket does.

T2-12

24Is there a viable entry point below national-brand terms?

Benchmark

Micro tenancies of 20 to 50 sqm, pop-up spaces around 25 sqm.

Fails when

Every tenancy is sized and priced for a national brand, and the only path in is a five-year lease with a bank guarantee. The independents are what make the place worth choosing and they cannot carry national-brand fitout and rent structures.

T2-13

25Have the eight daily economies been mapped, and do the leases require the hours the strategy depends on?

Benchmark

Morning and wellbeing 6 to 9am, breakfast 7 to 10am, transit at the peaks, lunch 12 to 2pm, afternoon and care 3 to 6pm, dinner 5.30 to 8.30pm, night-time 7 to 11pm, and true late-night after 10pm, all resting on a wellbeing layer of walking, exercise, caregiving and social connection.

Fails when

The evening economy is a strategy document position and every lease permits closing at 5pm. A precinct that only works at its peak is failing the other twenty hours. Night trade also needs a physical home: it concentrates or it does not happen. And the mapping has to show overlap rather than adjacency. Value is created where the same space serves different people in different ways at different times, not where a gym and a cafe happen to share a wall.

T2-14

26Is one party accountable for the whole precinct, and is that accountability funded?

Fails when

Each landowner manages their own asset and nobody owns the ground plane between them. This is the most common reason a well-conceived precinct degrades within five years. The anchor creates the halo, the halo creates the precinct, and the precinct only works when someone is responsible for the whole thing rather than the building in the middle.

T2-15

27Is the publicly accessible space governed and funded by a written instrument?

Benchmark

A Publicly Accessible Areas Management Plan agreed with the council, on the model used at Rouse Hill Town Centre, where public space management and programming is funded from centre leasing income. Australia Square, Central Park Chippendale and South Eveleigh are the other defensible Australian precedents.

Fails when

Public access is a condition of consent with no funding model and no management plan attached. The objections to privately owned public space, restricted access, restricted protest and photography, are all answerable through the management plan. They are not answerable afterwards.

T2-16

28Is there a cross-subsidy mechanism that protects the operators who created the identity?

Fails when

The precinct succeeds and prices out the independents, makers and cultural tenants whose presence was the reason anyone came. Without a deliberate mechanism, success is the failure mode. Collingwood Yards is the working Australian model.

T2-17

29Has activation been sized to demand rather than to floor area?

Fails when

An activation mandate is written against square metres. Activating too much space with too little demand disperses energy so thinly that nothing earns the trip, and produces the curated sterility that reads as busy in a report and empty on the ground.

T2-18

30Has Country been engaged, and is it recorded who was engaged and when?

Fails when

Connecting with Country is delivered as an artwork commission at the end of the design process. It is a design input, it belongs at the start, and the engagement itself is the part that has to be real.

T2-19

31Has anyone decided what will be measured, and who holds the number?

Fails when

Performance is read from the rent roll alone. Count pedestrians, dwell, vacancy, turnover by category and the ordinary Tuesday. What is not counted is not managed, and the instrument has to be set up before opening to have a baseline.

Tier three: the fourteen you can fix on a Tuesday

These matter. They are also cheap and fast to change, which is exactly why design review is the wrong place to spend the hour on them.

T3-01

32Is there enough seating, and is it moveable?

Benchmark

Moveable seating is associated with dwell time increases of 30 to 50 per cent (Gehl). Roughly ten moveable chairs cost about the same as one fixed bench (PPS).

Fails when

Seating is fixed, bolted and arranged facing the traffic. Moveable furniture is a philosophy rather than a furniture choice. Flexibility signals trust, and trust creates ownership.

T3-02

33Can a person sit down without buying something?

Fails when

All the seating belongs to tenants. A ground plane where the only place to rest is a paying seat will not hold parents, older people or teenagers, and those three groups are a large share of weekday presence. If there is nowhere for a teenager to be, they gather where they are unwelcome and the response becomes a security problem rather than a design one.

T3-03

34Will the landscape survive?

Benchmark

Species chosen for the light and water they will actually receive, with native and low-water species considered.

Fails when

Planting is specified for the render and replaced twice in the first three years.

T3-04

35Is the place legible across the day?

Fails when

At 7pm half the tenancies are dark, and a visitor cannot tell from the entry whether anything is open. Lighting, signage and clustering should make the answer obvious from 30 metres away.

T3-05

36Are there daily rituals the place supports, not just events it hosts?

Fails when

Activity depends entirely on programming. Rituals are the coffee on the way to the station and the walk after dinner. They cost nothing and they are what the ordinary Tuesday runs on.

T3-06

37Is there a programming calendar that matches the daily economies, and is it funded as an operating cost?

Benchmark

Bryant Park Corporation generates approximately USD 30 million a year in private revenue with no recourse to city funding (Bryant Park Corporation FY2024 audited financials, KPMG).

Fails when

The calendar is four large events a year, funded from a marketing line that gets cut in the first difficult year. Four events do not build a habit. A weekly rhythm does, and programming is infrastructure with a running cost.

T3-07

38Are toilets, parents rooms and drinking water planned for comfort rather than compliance?

Fails when

Amenities are placed at the end of an unlit service corridor. People remember the toilets.

T3-08

39Is safety designed in through activity rather than added through hardware?

Fails when

The response to a dark corner is a camera. Passive surveillance, lighting, activated edges and residential overlook do more. Busy streets are safer streets, and safety compounds into more use.

T3-09

40Has someone walked the whole route for accessibility, not just checked the compliant path?

Fails when

Accessibility is a ramp somewhere. Walk it with a pram, a wheelchair and a trolley, and check that facilities are identifiable for people with vision or hearing impairment.

T3-10

41Is arrival by every mode resolved on the ground: on foot, by bike, by bus, by train, by car?

Benchmark

Pedestrians spend up to six times more than car drivers, and footfall rises 20 to 35 per cent on well-managed shared streets (Living Streets UK, 2024).

Fails when

The car arrival is designed in detail and the bus stop is 200 metres away across a road with no crossing, with no bike parking anywhere.

T3-11

42Is parking sufficient, and is it a decent experience?

Fails when

Parking is either treated as the enemy or treated as the answer. Successful high streets have sufficient parking and slow traffic, not no traffic.

T3-12

43Are there design guidelines, a style guide, and a standard for hoardings and vacant tenancies?

Fails when

Shopfront signage is negotiated tenant by tenant and the street loses coherence in 18 months. Vacant tenancies with papered windows do more damage than the vacancy itself.

T3-13

44Is there a tenant community, and a review cadence with the original assumptions in the room?

Fails when

The only contact between management and operators is the rent notice, and the strategy is filed at practical completion. The assumptions written down before opening are the test plan for the first three years of trading.

T3-14

45Are operating costs under control from the start?

Benchmark

Energy typically runs at 2 to 5 per cent of tenant turnover, and savings of around 30 per cent are achievable through design and systems.

Fails when

Energy performance is a sustainability chapter rather than a line in the tenant's profit and loss. It is the latter that determines whether they can pay the rent.

Your assumption register

Everything you marked assumed. These are the things this scheme is currently betting on. Each one needs a test and a date before capital commits.

Nothing marked assumed yet. As you work through the checklist, anything you cannot point to a written answer for lands here.

What this checklist deliberately leaves out

It does not score. A number out of 100 invites the precinct to optimise for the number, and the number would be dominated by tier three, because tier three is where the easy wins are.

It does not cover the items that belonged to a different era of asset management: printed directories, designated smoking areas, customer service desks, window displays lit after hours. Some of these mattered in an enclosed centre. They are not what determines whether a ground plane trades.

It does not replace the strategy. A checklist tells you whether the answers are present. It cannot tell you whether they are the right answers for this place. That work happens before the list comes out.

The test that matters

Every question here is a version of one question. It is not asked at the opening, and it is not asked in the render.

A ground plane is judged at 11.40 on a Tuesday, by someone who had somewhere else they could have gone.

Related reading

  • The Street-Width Monogram and the Three-Coffee Walk, in Laneways, Widths and What Works
  • Streets That Trade: the shared street evidence base
  • The commercial case for privately owned public space
  • The Anchor Halo: governance, cross-subsidy and the halo
  • The 18-Hour Day and the overlapping daily economies
  • The Everyday Needs Benchmarks
  • The Retail Design Blueprint, seven pillars
  • Rethinking Public Spaces, six-part monogram
  • Flexibility and Trust in Public Spaces
  • When Activated Means Empty
  • There Is No Silver Bullet

Frequently asked questions

How is the checklist organised?
Forty-five checks, sorted into three tiers by how hard they are to undo. Tier one is the twelve you cannot undo, locked by the built form. Tier two is the nineteen the deal locks, set by the mix, the leases and the rent model. Tier three is the fourteen you can fix on a Tuesday.

How do I mark each item?
Three ways. Resolved means there is a written answer somewhere you could point to. Assumed means you think you know but nobody has tested it. Not applicable is a legitimate answer, used deliberately.

What is the assumption register?
It is the list the checklist builds automatically from everything you marked assumed: the things the scheme is currently betting on. Each one needs a test and a date before capital commits.

Is my data saved or sent anywhere?
Your marks stay in this browser. Nothing is sent to Bellringer unless you choose to email yourself the marked-up checklist and assumption register.

What is this checklist for?
Mixed-use, retail-led ground planes: town centres, high streets, build-to-rent, mixed-use precincts and transit-oriented development. Health, education and workplace precincts run on the same methodology but need a different subset.

We will email you your marked-up checklist and your assumption register. If you tick the box below we will also send new Bellringer briefs, roughly monthly. Nothing else, and you can stop it in one click.