Please reload

Recent Posts

Ideas on AI applications in the residential property development sector

April 7, 2020

Introduction – full article reading time 10 min 

This article is based on a 2 stage thinkaton that took place under ERP & Friends leadership in Feb-Mar 2020. The purpose of the thinkaton was to develop a view of the benefits of using AI in different applications for property developers. The overarching target was to find uses that would allow developers to strengthen their capability for profitable growth.  

 

 

The thinkaton framework we designed consisted of the following perspectives: 

 

  • Profitability in sector - Cost development history in the industry and the cost structure of properties built 

  • Product-market fit for different actors.

  • Competitive environment.

  • The AI capability must relate to the business platform 

 

 

 

 

 

 

 

 

 

 

 

Based on the above framework three strategic focus areas were identified in the first step of the challenge, these turned out to be:

 

A. Customer relationships 

B. Performance management

C. Procurement

 

Although we used a fictitious company with a certain profile we believe the findings in this paper are general for the industry. 

 

The property development sector and our “client” Lyan

 

 

Over the years the property development sector has seen cost developments in excess of the general CPI development in Sweden.

 

In the chart below from “Sveriges Byggindustrier” the factor cost developments are shown indicating that Material, Transportation and OH have been key drivers behind the increases. 

 

 

 

 

 

 

If we instead look into the cost factors that make up an average property project it becomes clear that three cost containment strategies may be the most influential on future profitability: 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  1. Procurement of land 

  2. Procurement of materials 

  3. Performance management of people 

 

Taken together these 3 make up 60% of cost. 

 

When looking at costs we risk losing sight of product pricing and what supports a strong relative pricing to the competition and the product-market fit. We cover that below. 

 

Product – market fit 

 

 

In our thinkaton we decided that our fictitious company Lyan was selling high-end apartments in attractive locations thereby being able to charge premium prices. However, attractive locations also mean higher than average land purchase costs. Our idea about the market character and Lyan is shown in the graph to the left, using the “Blue ocean” way of picturing the market. However, Lyan in somewhat of a “Red ocean” by any standards.

 

 

Further, Lyan is actively working with lifestyle desires of its customers meaning it needs to know its target customers more intimately than the competition. As will be evident in the application of AI later on, this was recognized in the thinkaton as a key success factor in beating the competition and being able to price it product with a premium. 

 

The competitive environment 

The property development sectors has certain aspects to take into account. We choose for our thinkaton to use the Porter framework. This analysis told us the following: 

  1. Suppliers – land prices are many times set by municipalities, unions are limits performance improvement for fear of losing jobs, some material suppliers maintain oligopoly making price negotiations hard, large projects requires large subcontractors, un-bundling of material+service is sometimes difficult to do 

  2. Customers – already explained to some extent under product – market fit above, in addition the level of mortgage rates heavily influence pricing power for developers and usually, in Sweden, supply is less than demand 

  3. Barriers to entry – you typically need a lot of equity and good relations with banks to finance projects, you obviously need domain specific competence such as construction estimating and procurement, you need steering/core/support processes to run the business and track-record to be trusted by partners 

  4. Substitutes – Customers may rent, buy on the secondary market, build on their own, have less m2 per household and person and so forth leading to a ceiling on pricing per m2. 

The new capabilities for future profitable growth 

In the idea generating stage many more ideas were generated so what follows we discuss the most promising ideas that the team found. This is considered a discussion starter with the industry so please get back to us with your thoughts. We would love to get your take and jointly formulate a view of of the future. 

 

Capability 1 – Land purchase prediction 

Capability 2 – Automated delivery and quality measurement using embedded rfid and IoT sensors 

Capability 3 – Matching engine for finding the right future competence and talent 

 

 

Capability 1 – Land purchase prediction 

The purpose of this capability is to bring down the cost for the 20% of project costs that land make up according to the project cost factors shown earlier. We think that by using AI algorithms it will be possible to forecast the best land purchases using a much wider dataset than used today. In particular we see the availability of Big Data sources to allow the collection of a wide set of data. In particular we want to find determinants of future location attractiveness not obvious to the market today.  

 

This will give us an edge in land purchases and since a lot of the profits in property development are generated by buying the right location and match the product with the location we believe this capability is super-attractive to developers. Just imagine getting land a couple of percentage cheaper for each project and the business case value of that.  

 

A secondary effect of this algorithm is that land valuations can be made more accurate by valuation consultants, however, this was not our primary prupose of this capability. The details of this capability can be found in the appendix.  

 

Capability 2 – Automated delivery and quality measurement using embedded RFID and IoT sensors 

First, using RFID, passive or active, is not that new. What we believe is new is the price drop and hence the feasibility in requiring all material deliveries to be tagged. This entails quite a change to the industry as each company would need to supply their own tags to the supplier to ensure full tagging. Until all suppliers are including tagging with each delivery. 

 

With active tags it will be possible to track deliveries real-time but the cost is still too high to be an option for all material. Rather we see a mix of passive and active tags being used. 

So why have the tags? The purpose of this capability is to allow automatic tracking and goods-received-on-site and semi-automatic quality control and waste. The benefit of this capability is to reduce disruption to production, visibility on project procurement progress and to reduce waste. Further, with time, the AI algorithm using the data generated will predict delays and waste levels based on location, project and suppliers combinations. And ultimately, the algorithms should be extended to suggest changes to project and supplier combinations to mitigate delay and waste risks. We believe this will be a quite strong tool both for owners, designers and project managers.  

 

As can be inferred from the above, a lot of new data needs to be gathered to feed algorithms with and that of course is something that could start right away. What master data model do companies have today to classify projects? Many companies limits the set of master data collected because the value of spending time collecting it is not clear. Now it will be the other way around, how do you motivate not collecting more data about each project? The structure of your information model is becoming increasingly important. It is time to move on it! 

 

Capability 3 – Matching engine for finding the right future competence and talent 

As hours spent building a certain building makes up 20% we wanted to reduce waste in hours not being productive. There are studies showing various extent of non-productive time on site. All of them show massive savings potential. We believe the contractual relationships between various actors in the industry; between different labor disciplines, between project management and laborers, between contractor and subcontractor and finally between contractor and developer makes performance management and cost containment an extremely complex problem to solve. And not a problem that necessarily technology will solve but people. 

 

So, we decided that getting the right competence given the future demands of the business would be something a technology could support companies with. Also here we see the availability of social media data such as LinkedIn, Facebook and others maybe used to both extract information about candidate search patterns and candidate profiles. Over time the data collected will be possible to use to train employee-recruitment fit algorithms from the vast amount of candidates in the market. We think the value of this will be even more profound as the future roles in a property development company will be different than today. Imagine knowing who is likely to be approachable with a job offer just when that person has started to think about a switch. And that it is the best fit with your company as well. This way company performance will gradually build up as your company fetches the best competence for the job! 

Summary 

 

The will be huge potential going forward in using the big databases that are growing as we speak. Correctly used and with the right technology they will support better decisions that ultimately will make every company perform at a higher level. In the three capabilities presented in this paper it is also possible to start experimenting with them without disturbing day to day operations. So, are you ready to do it? 

 

Let us know if you would like us to support you in this process, we are more than happy help in whatever ways you find useful. Further, if you would like the more detailed descriptions for each capability let us know, we are happy share! 

Share on Facebook
Share on Twitter
Please reload

Please reload

Search By Tags
Please reload

Archive
Erp & Friends logo

© 2017 by ERP & Friends AB

  • Facebook Social Ikon
  • Linkedin Social Ikon

Contact

stefan.schnitzer@erpandfriends.se

håkan.lundmark@erpandfriends.se

RSS Feed