This is a guest post by Anita Ngai, who has extensive experience in technology, retail and urban development. She worked in McKinsey for 4 years before transitioning to Real Estate in Hong Kong, and online travel. She was trained as a structural engineer. She is currently exploring a start-up idea focused on helping developers become more data-driven in their planning and leasing processes. You can contact her here.
I love that domain experts – in this case – a structural engineer cum real estate professional, are thinking about how technology can transform the way their industry works. Hope you enjoy her article as much as I did!
WHY IS THE MALL DYING?
“The Death of the American Mall”, “Ghost Malls in China”, “Are Malls Over?”, “Is the Physical Shopping Mall Dead?”, “China’s Ghost Towns and Phantom Malls” – if you google search the term “shopping malls”, these headlines pop up. What’s interesting is that these headlines cover places as diverse as the Midwest US to the large metropolitans of China. Some reasons for this trend include:
- Online shopping
- Urbanization – higher concentration of population and/or wealth means less retail space needed in suburbs
- Changing demographics – deceleration of population growth, aging core group
- Slowing income growth/increasing inequality – weaker GDP growth; wealth more concentrated in hands of a smaller number of people
- Change in consumer preferences – trend that millennials prefer to live and occupy less space
- Overbuilding catch up – we have been overbuilding for some time, and it’s finally catching up (as New York Times quoted a real estate executive: “The mall genie was out of the bottle, and it was never going to come back.”)
- Poor management – bifurcation of malls into great versus terrible ones that don’t survive
The death of the malls poses serious challenges to developers and planners. Their previous paradigm, “build a mall and people will come”, no longer holds today. Instead of building new malls, developers need to focus on conversions and repurposing of existing malls and spaces.
THE PROBLEM OF UNDER-UTILIZED ASSETS IN A CITY: WHY DOES THIS MATTER?
Underutilized mall spaces are not just a problem for developers – they are a waste of a city’s precious land resources. For example, in dense cities like Hong Kong, where I worked in real estate in different roles for four years, the competing demands on land are very real – retail is very much in demand by the upper-middle class and mainland Chinese tourists. On the other hand, the housing crisis is getting more and more acute because of the lack of space for new housing developments.
Instead of allowing new retail spaces to be built nearby, or even tearing down retail spaces, it makes more sense to convert and enhance existing retail spaces. Maintaining density levels in urban and suburban areas can bring socioeconomic benefits. Furthermore, the carbon footprint of retrofitting has been shown to generally be much lower than demolition and rebuild. All this means that potential public and private investments into our built environment can be better directed, to projects with higher value to society.
THE PROMISE OF TECHNOLOGY IN BOOSTING MALL UTILISATION
In the age of Airbnb and Uber, one would think we could do better in optimizing the underutilized assets in malls. Indeed, technology holds tremendous potential in helping developers do this – both at the planning and the post-completion stages.
Collecting and analyzing data can help developers customize their projects to their potential users. In the past, developers only had blunt demographic data (population size, income levels, age composition) on which to base their plans. Now, sensors and mobile phones can capture large volumes of finer data e.g. what types of shops women between 30-40 in the geographical vicinity dwell longer and spend more money at.
Combining all this data, developers can use sophisticated statistical simulations and machine learning to predict the foot traffic, occupancy levels, and likely visitor profile (e.g. income-level) of the project if they vary the proportion of space dedicated to retail vs entertainment vs hospitality/accommodations.
Testing hundreds of scenarios of the project mix and layout would only take seconds, but is close to impossible for humans to do – both from data collection and computational analysis perspectives.
After the project is built, there are decisions that developers and their leasing teams have to make continually – who should we lease each space to? How should we price each space? How long should the lease period be for each space ( the default now is 3-5 years depending on the market which works for some, but not for others).
Each of these decisions has tremendous ramifications for the mall’s utilization. For example, putting a fast-food restaurant at a certain entrance to a mall would draw a lot more footfall through that door, versus a beauty supply store or the front lobby of a three-star hotel. For each space in a mall – whether a back corner on the ground floor or center core on the third floor, a fast-fashion tenant, quick service restaurant or three to four food court stalls will each have a different footfall impact, chance of success, and likelihood of sustaining their business over the long-run.
Developers also need to be more flexible with the use of space – pop-up stores, for example, have helped ease some of the long-term vacancies or low footfall issues that landlords are seeing in their retail properties. But this is not done in a data-driven or widespread way: pop-up stores are often under the purview of marketing teams, and theleasing teams may only take a support role.
If developers collect and analyze data effectively, they will also be able to lease their spaces and re-configure their malls based on real-time data. All this boosts utilization and uses space most efficiently.
SO WHY IS IT NOT HAPPENING? AN INSIDER’S PERSPECTIVE
Having worked in real estate for a number of years, here are the factors that hold back these obvious innovations from taking off.
The first reason lies in how developers think about innovation. The only teams within those organizations thinking about innovation and technology – some form of a “digital” department and an incubator/VC – are not usually tasked with looking at the design process. They focus on “downstream” issues like improving customer experiences in a shopping mall or on having a bet in a start-up who will “hit it big” one day.
Second, even if a developer/owner is motivated to take a data-driven approach to design, a single company’s portfolio of property may not be large enough to yield data that is representative of the market view. Certain Asian developers come closest to controlling the ownership of an entire neighborhood or district, but worried about competition, they would not be motivated to share this data with the industry, competitors or brokerage firms.
A third reason is similar to what we have seen in many other industries: existing players will only make incremental changes, until someone new comes in to disrupt traditional practices. Tech start-ups have been active in the real-estate sector, but mainly in three areas:
- Real estate transactions
- IoT and smart homes/buildings/cities (the fridge that will order for you when you’re out of milk, the trash can that sends a signal when it’s full and needs to be serviced) and
- Visualization (VR for potential buyers to walk through their unbuilt/faraway home; 3D rendering and VR experience of construction blueprints).
Unfortunately, I have not seen many start-ups work on applications that will help with the design and planning of malls. There are a few providing heat maps of where footfall is in a mall; or analyzing the type of store a given neighborhood needs, e.g. apparel, doctor’s office. Mapping start-ups are currently focused on other areas of applications, such as self-driving cars.
MY HOPES FOR THIS SECTOR
Retail makes up a significant portion of a city’s built space inventory: San Francisco has about 76.3 million square feet of office space versus 80.5 million square feet of retail space. It will remain a useful and desired part of city life for time to come. However, it will be a costly waste of precious city space if the trend of underutilization continues. Developers will be able to buck this trend if they use a far more data-driven approach to planning and leasing.
I sketched out the challenges above, and I believe they can be overcome if developers can take a longer-term view to invest in evolving their planning and design processes and to incorporate new data and technologies available. The benefits from using new approaches are not easily quantifiable without having tested them, so sticking strictly to ROI figures will not lead decision makers down this path.
Also, more startups and public agency collaborations such as Uber Movement and World Bank’s Open Transport Partnership would allow the immense amount of data being accumulated to become transparent for public use. Having a public agency host data from different private sources may help overcome more data privacy concerns floating around, though these agencies would likely need tech companies to help them improve on data security. Governments can play a more proactive role in facilitating progress, through regulations and test projects, and I believe the municipal level – because of smaller size and relatively less partisan impasse – will be the best testing grounds.
 (Of course, there are still places where there is a real growth in the population or local economy, and so new retail space is indeed needed.)
A number of studies actually show that higher densities can lead to higher public expenditure per capita, though there is evidence that this is due to government management practices, e.g. higher government employee compensation. In addition, lower densities do not necessarily increase public expenditure because the costs for sewage, electricity and other infrastructure are actually priced into the new houses, i.e. bore by the residents themselves. Benefits from higher density developments are more obvious if we include quality of life metrics (e.g. traffic congestion, air pollution).