Customer Science in Motor Retail: Keynote 2. How to Harvest Actionable Insights from Dealer Management Systems

Most motor retailers employ a Dealer Management System (DMS) to manage and administer all facets of their business operations.  The typical components of a Dealer Management System include Financials, Showroom and Customer Relationship Management (CRM), Parts, Service, Workshop Management, Point-of-Sale, Franchise Data, Vehicle Stock, Purchasing and Pricing Administration.Actionable Insights are those views of data that cause managers to ask new questions about how processes work and take action.  They differ from key performance measures and daily operating control (DOC) reports that focus on delivering a picture of progress against a strategic objective, operating budget or forecast.  What actionable insights can dealers expect to source from their Dealer Management Systems – and how do they source them?

...

Many of the Dealer Management Systems that exist in the motor retail sector are proprietary systems and some of the technology employed dates back to the pre-noughties when the world had never heard of cloud computing and mobile phones were the size of a small brick.

The reason why the industry has been so slow to adopt change in IT has a lot to do with the complexity of the dealer information environment and the close synergies of the many operational areas that exist in a dealership.  As part of a tightly knit supply chain, dealers also need their systems to talk to the systems operated by the motor manufacturers.  These integrations are also complex, hard to create and adapt. Dealers ideally want to have a single information environment so to meet their need. This means software vendors have to invest hugely in attaining the ‘critical mass’ of capabilities the industry demands before they can hope to compete with incumbents.  This creates a natural barrier to new software vendor entrants hoping to enter the market.

The consequence of this market stagnation in viable competitive DMS options for major dealer groups means many are required to operate computer systems that are poor at sharing their data and exposing the actionable insights that dealer execs need.

In a market where competition is growing daily, dealers know they need to be much better at understanding the relationship with their customers, better at producing personalized offers that are event-driven, fine-grained and relevant, and better at finding more reasons to speak to customers that are ever more distant and only happy to engage with suppliers when it suits them.

Data Discovery from Dealer Management Systems

The quality and completeness of operational analytical tools found in Dealer Management Systems varies hugely.  Many of the younger software products include highly impressive reporting tools that are ideal for understanding the performance of pipelines and processes.  Where most dealer systems tend to fall down is their ability to source ‘holistic’ views of customer profiling, interaction and life-time value.

The essence of customer science is to gain a single page view of the customer that uncovers all aspects of the relationship including:

• Profile: Affluence, location, family status, employment, habits etc.
• Persona: What makes them ‘individual’
• Communication: Communications history, activities and preferences
• Driver behavior: How the customer drives their vehicle and the impact on vehicle depreciation and maintenance schedules
• Buying behavior: What do they buy, how do they buy, likes and dislikes; what sort of relationship do they want to have with their suppliers?
• Products: What vehicles do they have, like, have had previously?  What other products and services have they used, would they like?
• Life-time value: What is the potential value of business derived from this individual

Uncovering these insights is valuable – useful - but it’s only when these many attributes can be connected that the eureka moments occur and marketers and their customer service colleagues uncover the gold dust that makes their campaigns more effective and raises the bar on their customer experience.

As a general rule, Dealer Management Systems are designed to support the many key processes that exist within a motor dealership.  By analyzing generic capabilities of motor dealerships we now know there are approximately 136 major processes that occur within a motor dealership.  These are supplemented by close to 200 common reports and approximately 150 primary referencing tables.  You get the picture – it’s complicated.

When software applications are built for a purpose the data models that underpin them take on that form.  In the case of Dealer Management Systems that means that most data models are poorly designed and include the same tables time and again in different parts of the data structure for different reasons.  For example, the field ‘Country’ will appear in customer and supplier addresses.  It will also appear in delivery locations and other records.  But will there be a single table for ‘Countries’?  Probably not.

So what?

The implication of this confused data structure is that execs that want to see customer views by activity, service used, spend (etc.) won’t be able to easily source these views because the data will be held in many separate silos and in the wrong structures too.  It also means that dealers, when they look at the data, they won’t necessarily know if they’re examining a list of vehicles from the CRM system, or a similar list of vehicles from their Vehicle Stock Book.

Advances in ‘Actionable Insights’ Data Harvesting

The good news is that cloud computing has introduced smarter ways of harvesting and creating new views of data that remove the hit and miss complexity of constantly trying to mash views on a spreadsheet.  How it basically works is this:

1. Flat files reports (the type you produce on a spreadsheet with no pivot tables) are produced for each master process area.  These are designed to contain the magic fields that identify customers, vehicles, sales people, dealerships and other key identifiers.
2. These flat files are uploaded to a ‘sorting’ engine that normalizes the data, uncovers errors and brings the various flat files together in the form of a relational database ‘in the cloud’.  This new data structure is sometimes called a ‘data mart’ – not to be confused with a ‘data warehouse’ which is a much more rigid, pre-planned data architecture for preprocessing views of data. 
3. Once the data is onboarded, data connections are automatically generated between the variousmagic fields (i.e. primary key identifiers of the subject records) and it’s now possible to start performing queries on the data to create the actionable insights execs want to see.

The above summary ignores the fact that there’s a lot of data NOT on Dealer Management Systems that can enrich customer profiling systems and add fresh perspectives to analysis such as customer experience data, website clicks and interactions, finance company data, service plan data, SMS messaging inbound/outbound activity, mapping resources and consumer profiling data resources from third party vendors.

Federated operational insights systems like the one I describe are best placed to harvest data in this way and remove the complexity and cost of sourcing data from Dealer Management Systems.  A few words of warning however:

• This type of system does not displace the need for effective operational reporting within the process-centric tools that run the business, though these are increasingly well catered for by vendors
• Operational analytical systems are rarely able to feed data back into the Dealer Management System, although why you’d need to do this I struggle to understand.  I suspect most managers want a ‘single version of the truth’ for the data they look at but few people care if it resides on one system nowadays.

Key Technology Components

The good news is that cloud computing has introduced smarter ways of harvesting and creating new views of data that don’t require people to invest time manually authoring reports or playing with data in spreadsheets – and thank goodness because anyone that’s had to do it knows how boring and tiresome to task of manually aggregating and normalizing data can be! 

There are a few crucial technology innovations that have made this possible:

Web Server Apps

Popular platforms like the Microsoft Web Platform (comprising of modules including Microsoft SQL Server, Microsoft ASP.NET and Microsoft IIS) are able to publish applications in a format that means they can be viewed on a browser without requiring any downloads and plug-ins to the computer device.  That means you can view applications using mobile phones, tablets, PCs, laptops and smart digital televisions provided you have suitable User Permissions to login. Users are assigned to User Groups and their access and usability permissions are governed by the permissions assigned to the group, or groups, they belong to.  In a 24/7 world, users can experience the Martini moment of – anytime, anyplace and anywhere computing without having to fuss with onboard apps.

Private-Cloud Infrastructure

Deploying hosting web server apps on a private-cloud infrastructure means that buyers don’t need to fuss about installing web servers and managing them.  It removes the burden of installing hardware and software applications; and managing them.  Private-clouds can be configured for each account (so that each account has their own dedicated database and resources), or alternatively clouds can be multi-tenant environments where customers share the computing resources of the host web server.

Data Flow Augmentation and Data Extract, Transform and Load (ETL)

Something has to engineer the automated uploading, normalization, transformation and workflow of data to the cloud environment.  The tooling to do this now is relatively common-place.  What this automated routine needs to do is to watch a folder and wait for reports to turn up, or automatically kick into life at a scheduled time, to then audit the content and coax the data into the technology equivalent of a food blender that will spit out ‘good data that fulfils the upload criteria’. 

One of the clever features of this ‘blob’ of technology is that it needs to be capable of dealing with iterative changes to successive uploads of the same reports.  Let me have a go at explaining this:

When you upload several reports to the data crunching engine in the cloud for the first time, all is good with the world because there’s nothing else up there.  Do it again and the data crunching engine needs to sift through the fresh upload of data and work out: (1) What data exists and hasn’t changed? (2) What data exists but has been updated? and (3) What new records have been added that need to be included?

This is a non-trivial challenge and is an important feature of the data crunching engine.  Otherwise, the next time data is uploaded it will result in duplicates and lots of confusion (not good!).

Operational Analytics Software

There are many different sorts of operational analytics software (see my summary article on http://www.business2community.com/business-intelligence/operational-analytics-best-software-for-sourcing-actionable-insights-0560328 for more details) but some kind of data visualization and reporting tool-kit is needed to make the most of the harvested data.

Combine these four attributes – web apps, private-cloud deployment, data flow augmentation/ETL software and operational analytics software – and you have the essential ingredients of an Actionable Insights system that can harvest data from your DMS without have to work hard to do it!

Purchasing Options

There is a wide spectrum of technology vendors able to provide either components or complete platforms for authoring operational insights systems capable of extracting data painlessly from a DMS. There are also companies out there that provide marketing, data cleansing and customer science services targeted towards serving the specific needs of the dealer market. 

When it comes to the computing platform itself, as I summarize above, any solution needs to manage data crunching (to get the data in the right shape and form) together with data visualization and reporting tooling to make sense of it.  These ingredients with then need to reside on a server somewhere; more commonly now on a private-cloud to prevent the need to manually install hardware and software.

Buyers have a range of sourcing options to consider and these include:

1. Build your own system by using a highly dexterous and enterprise scalable cloud platform

There are too many of these to cover all of them but examples include the obvious BIG names likeMicrosoft Azure, cloud.google.com/appengine, Force.com (from Salesforce.com), WebSphereClouds orAmazon Cloud.  These big cloud tool-kits make development much easier than it used to be and all of the major vendors are on-point when it comes to scalability, security and resilience.  Don’t expect to have a quick solution though because it takes time and effort to get solutions ‘tuned’ to suit your needs.  Nevertheless, once you’re done you should have a first-class system.

In the last few years we’ve also seen a proliferation of smaller, expert cloud app platform vendors likeCloudFoundry, CloudAppStudio and ActiveState that make it easier to design and publish custom applications to secure cloud environments, often demanding lower skills overheads.  The challenge facing these vendors is to ensure they can offer sufficiently complete platform capabilities.  They need to satisfy customers that they’re able to go the last mile on developments: The last thing customers want is to find, after months of a development, that some capabilities are missing and they have to start over on a bigger platform like IBM WebSphere, Amazon Cloud Web Services or Microsoft Azure that profit from the many millions of R&D dollars that bigger vendors have at their disposal.

2. Buy breed tool-kits

Getting a solution up-and-running faster can be achieved by selecting best-of-breed components. Two will be necessary to ‘fast-track your development;

1. A platform for designing the apps you need and managing the cloud data crunching using platforms like Interneer, Encanvas and OutSystems that offer out-of-the-box tooling to meet the majority of needs.  All of these providers provide ‘low or no’ coding overhead to the task of authoring applications.  They also possess rich data connectivity and workflow features.
2. An overlaid expert analytics tool-kit like Qlikview, Tibco Spotfire, Tableau, Yellowfin, Birst,iDashboards, Jedox and Jaspersoft.  These providers offer data visualization, dashboarding and reporting tools.

Of course, the nature of competition means that if you ask any of these vendors they will probably say they can offer a complete solution ;-)

3. Pay someone to build it for you

There are many companies out there like CSC, CACI, Canon, US Tech Solutions, Wipro, TCS, PA Consulting and countless smaller industry expert companies like DealerSolutions, Nybble and Ambridge Consulting that specialize in the data centric custom applications authoring arena.  These companies will provide fixed price projects to deliver solutions to buyer specifications.

4. Buy a ready-made service

I have no doubt that, like NDMC Consulting and introhive, there are other companies have entered the automotive market in the past few years thanks to the available of suitable tools to provide ready-made online services that deliver the outcomes needed without requiring dealers and motor manufacturers to build their own solutions.  The challenge facing providers is that each and every system will require heavy customization and tailoring.  Encanvas Remote[Spaces] is an example of a cloud architecture that enables every customer to enjoy their own individual and private cloud service.  Competing platforms like Interneer,Tibco Spotfire, Force.com and OutSystems (to mention a few) are no doubt able to offer a similar capability.

Summary

Cloud-based operational analytics solutions are growing in popularity as a way to extend the life of Dealer Management Systems investments.  They lever value from sleepy data held in administrative systems by finding smarter ways to turn data connections into new reasons to engage customers in dialogue.  These solutions do create a single version of the truth but in a different way; not by seeking to install another system, but instead by adding a value extracting capability in the cloud that gives dealer executives the actionable insights they need without the pain of yet another major IT project.

About This Series of KeyNote Articles on Customer Science in Motor Retail

How does Motor Retail benefit from applying operational analytical tools to source actionable insights?  In this series of keynote articles I explore how the coming together of mobile computing, innovations in customer science and advances in big data analysis are creating the perfect storm for marketers in motor retail.

BIG DATA with Encanvas (infographic)

In this infographic I explain how Encanvas employs big data as its raw material to accelerate the pace of business innovation.  Of course, data doesn't have to be BIG to be useful.  Over the last 10-years I've seen many good ways of extracting value from data by simply having more data of different types accessible to business analysts. For example, one professional services company was able to improve their presentation of credentials by exploiting data from a third party database to enrich their customer records and map their customer records onto their project data to build a richer understanding of what they'd delivered (how and to whom) to improve the quality of their case stories.

Accelerating the pace of business innovation requires not just the means to harvest data and analyze it - businesses ALSO need the tools to improve their processes and embed learning lessons into their day-to-day activities.  This is by far no mean feat!

Encanvas overcomes the natural reticence of organizations to change, break age old 'norms of behavior' and embrace better ways of working by creating applications without asking organizations to invest in IT infrastructure or spend days, weeks or months on IT projects with no clear ROI.  Building a situational application in a day or half-day workshops session means that businesses can understand the value of data - and apply it - removing the risk and cost of process change.

Speaking to users of Encanvas, they will very often rebuild apps many times as they learn from their data and re-apply learning lessons to improve processes.  This would not be practical or affordable if a traditional software development or enterprise platform like Microsoft SharePoint was used to fashion applications.

The original infographic can be viewed on slideshare here.

Profit from BIG DATA: Using actionable insights to accelerate business innovation

Actionable insights are those that make managers and workers 'curious' and cause them to act. The big data revolution is giving organizations many more reasons to exploit data so they can improve customer experience, personalize offers, improve sales engagement methods, optimize processes and surface BLUE OCEAN opportunities.  Even your existing data that's locked away in back-office systems can become extremely value when combined with other data to answer new questions. But what PROCESS do you need to harvest actionable insights and act on them to accelerate the pace of innovation in your business?

In this article I've summarized the obvious 5-steps to turning sleepy data into (first) actionable insights and (then ultimately) better processes using big data cloud computing and operational analytics technology.  Then I've explained the 2-step process (yep just two steps) that Encanvas makes possible. While the example below is based on the capabilities of the Encanvas platform because we're familiar with it at NDMC there are other technologies from Tibco, IBM and Oracle that you could also consider to achieve the same outcome (though this is probably an 'IT' project if you do).

The reason we developed and continue to use Encanvas at NDMC is because it produces the fastest time to value by reducing the number of steps in the Accelerated Business Innovation process - and it deskills the activity of process innovation so you don't need to be an IT expert or programmer to implement it.

A typical 5-step innovation life-cycle process using a mix of enabling tools is:

1. Capture Data -  2. Hoard Data - 3. Harvest/ETL - 4. Analyze -  5. Act

This can be a long-winded and technically challenging process: It requires a different bit of software for every stage, and this results in large IT teams, expert skills and slow time-to-value projects.

Why 'Acting' on actionable insights is like exploring for oil
The accelerated business innovation process is modeled around oil exploration where you first make sure you've hit oil before you start to build infrastructure.  This means that when actionable insights surface a better way of working, the solution is to create a situational application (at little or no cost) to validate the process change.  If the new process works better and provides incremental value then it is adopted, if not, it is discarded.  That takes a new approach to IT and new tools. The most challenging part of adopting an accelerated business innovation process using legacy enterprise software is therefore the 'ACT' bit.  Adapting legacy platforms like Oracle, SAP and Microsoft Dynamics is more akin to turning an oil tanker - really difficult and expensive:  It requires IT people with deep skills.  That's why we use Encanvas to serve the long-tail of situational applications that are used to 'explore' value and then adopt those applications that prove valuable as a layer that sits above the fragile 'heavy-lifting' enterprise software that most businesses operate today.

1. Assimilate data

Data assimilation describes the task of bringing together data to a level of consistency and completeness that makes it useful.  The term means more than simply harvesting data - because anyone who has tried to bring data together from different sources (even using a spreadsheet) knows how time consuming and painful it can be to extract, organize and structure data.  Fortunately, tooling is available nowadays to automate many of these very manual tasks.

Many people assumption - probably due to the norms of behavior encouraged by decades of data warehousing - that it's necessary to hoard data and organize it before you can start to lever insights from it.  Nope.  Today, the data harvesting, data source integration, data augmentatation and normalization tools are so powerful that most operational analytics systems can 'get at the data' wherever it is and exploit it without having to first HOARD the data in some great big massive repository that you have to pay for.   (That said, if you do want to hoard data, technologies like Hadoop have made the task of hoarding data of different forms immeasurably less expensive and more scalable.)

Encanvas, and technologies like it incorporate a full range of data connection and ETL tools that obviate the need to code or use lots of APIs and middle-ware tools in order to harvest data.  It supports data integration with CCTV, security databases, hardware, sensors and telematics.  Whats more, it enables situational applications designers to exploit structured data (like relational databases, web services, CSV files etc.) and unstructured data (such as PDF documents). Their Information Flow Designer application gives applications designers the means to formalize data gathering workflows from data sources, transform, cleanse and normalize data - and then expose it for operational analytics applications like Encanvas Create Design Studio to exploit.

One thing we realized at NDMC a decade ago is that authoring exploratory (situational) applications to source actionable insights requires designers to play-around with data, forms and logic tools - all the toys needed to create exploratory apps in the same integrated design environment. That means you need the data sources to play around with too.  Encanvas allows us to try things out, get things wrong and work with stakeholders until we get a result.  Assimilation of data is tricky, sometime painful and rarely straight-forward!

2. Create and Deploy Situational Applications

Sourcing actionable insights normally means connecting to many different data silos and exposing the data held within them.  The 'clever bit' that technology can help you with is to bring data together in new ways, to then formalize business processes to act on it.

This 'iteration' of the enterprise computing environment requires to perpetual creation of situational applications that serve individuals or small communities of users that traditional IT can't service because this long-tail of demand is simply too long to be addressed by technology platforms that need coding - and users have such a diversity of requirements for tooling.  While there is a natural instinct to employ existing tool-kits like IBM WebSphere, Microsoft SharePoint or applications platforms like Force.com, art NDMC we foudn this to be a costly approach because it requires large IT teams and the tooling is limited by the capabilities of the platform. None of these enterprise platforms were engineered for the purpose of authoring situational applications and so IT people find themselves knee-deep in APIs, forms and database programming projects and third party data analytics and data source connectivity tools.

Using built for purpose environments like Encanvas, situational applications are created in a codeless integrated design environment. We use the same environment to author operational analytics as we do actionable applications - and very often an exploratory application for operational analytics can be progressively extended to become a key part of the enterprise operating environment.

In the normal course of an accelerated business innovation project, stakeholders will start by creating operational analytics applications to analyze their data.  Then they will create applications to ACT on the results of the 'data analysis' exercise.

When data does not exist, new applications are authored (for desktop, tablet or mobile devices) to capture data.  When authoring such applications it make sense to control the quality of data input wherever possible by adopting the use of drop-down choice fields, radar buttons,check boxes and check-lists. When working with Encanvas, even applications used to capture data can be 'situational'.  Very often these applications are eventually absorbed by 'platform' systems like Microsoft SharePoint, Oracle and SAP.

How long does it take to execute an accelerated business innovation project from 'concept to completion?
After a decade of examples we know that the answer depends on a number of factors:

1. The size of a project - Smaller projects (i.e. Something akin to displacing the use of spreadsheets for analysis of audit and accounting processes) can be actioned within a day, sometimes two.  Large projects take significantly longer.  When we've employed accelerated business innovation to upgrade a regional transport system, enhance an eLearning system, install federated risk or compliance across a business - that sort of thing - it normally takes around 6-weeks to get through the situational applications phase and move on to User Testing or 'embedding' the technology deployment into the enterprise platform.  While 6-weeks is a long time it is nothing compared to traditional IT projects that span many months, sometimes years.

2. The change methodology - We are supporting a change process and so inevitably HOW you run a project will impact on the pace of achieving outcomes. An NDMC we've created a Computer Aided Applications Development methodology to support accelerated business innovation projects.  This combines the 'good thinking' found in Outcome Driven Innovation (ODI) and Blue Ocean strategy mapping in addition to a few other methods that are unique to our approach.

3. The tools - If you're using a codeless and complete situational applications design environment like Encanvas Design Studio then 90% of the programming, testing and tuning overhead authoring process is normally removed from the exercise - and stakeholders can contribute to workshop-based developments.  This speeds up the innovation process because project leads can achieve an outcome during the course of a 3 or 4-hour session rather than just creating 'a list of things to complete'.  If you are required to use tool-kits like Microsoft SharePoint THEN YOU CAN still achieve great results but it does mean you will need a dozen more IT people to support the process.

I hope you found this article on accelerated business innovation interesting.  Do let me know of your own experiences.

Ian.

BIG DATA - The Killer-App for Motor Dealerships

BIG DATA is a big topic at the moment in IT... and that interest is gradually working its way down the corridor to the marketing office and the boardroom.

In the Motor Retail industry new IT innovations emerge somewhat slower than other verticals because many dealers rely on partner suppliers to manage their IT.  This creates a bit of a lag in adoption of new technologies.

I doubt it will be long however before the topic of BIG DATA filters into motor dealerships because the competitive advantage it brings is compelling.  Another reason why BIG DATA technologies are likely to take off so fast in this market is because there is pent up demand in dealerships to 'get moving' with innovation.  Many of the Dealer Management Systems employed today are WAY behind the curve on technology - some are not even based on web technologies.  The fragmented information environments that dealers have to cope with day to day are embarrassing to anyone in the computer industry (on behalf of all us nerds in the software in 'sorry guys').

At NDMC we're leading the charge to introduce BIG DATA solutions into the Motor Industry and the level of interest and take-up has been rather overwhelming.  In this presentation I've provided a short summary of the role and impact of BIG DATA for anyone who's yet to encounter the technology.

I hope you find it informative and a little fun! I.

BIG DATA - The Killer App for Motor Dealers from Ian Tomlin

Customer Science in the Motor Retail: Using Maps to Source Actionable Insights

http://www.drivingsales.com/blogs/CustomerSciences/2013/07/22/customer-sciences-motor-retail-keynote-1-actionable-insights-result-mapping-customers

This article provides some insight into the ways geo-spatial mapping can be employed in the motor retail market to source actionable insights that help to boost marketing effectiveness, reduce costs, enrich customer data and lost more...