What Appolis Tracks

Why does a company like Appolis put an ad in a parking garage with a QR code?  Well, this brings me to the topic of tracking.  In my analysis, most marketeers see the value in QR codes as providing tracking of customers and prospective buyers with targeted messages and instant feedback on “views”.  They engage the customer with an experience that can connect the physical locations of the codes with the message that is tailored for a consumer.

The dynamic nature of the technology also allows messages to change at any rate that is applicable to the experience.  If you scan a QR code in the morning, the result could be a recommendation of a breakfast cafe that is providing discounts.  In the evening, the message can be a movie review that may influence your decisions on what movie to see.

For Appolis, a company that has been built around mobile devices and tracking of mostly physical processes and assets, including food and medical devices, we asked our selves, how can a QR code be purposed for tracking inventory that is “off the grid”.

Enter the concept of remote inventory management.  About 2 1/2 years ago, Appolis invented a method to communicate inventory information using a consumers mobile phone.  The biggest challenge was that it required a user to enter numeric codes into their phone.  These codes contain encrypted data that represents a host of information about the inventory levels as well as an identity that only the receiving system could decrypt.  The problem was that these codes can only fit so much information in them.  By the time the code gets to 10 digits, good luck accurately entering them into the phone.  Why would someone take the time to enter a code they see on a vending machine?  Because they receive a reward such as free vended products.  QR codes are now a superior mechanism to get the “code” from the vending machine to the “mother ship”.

Appolis is bringing this technology and other consumer enabled solutions that allow ubiquitous tracking of not only physical movement and consumption of goods, but also, making it fun to use.

Check out more about what we track at www.appolis.com.

Android versus iPhone

At the beginning of the year, I handed my Windows Phone 7 over to a co-worker and decided to try out the latest Android devices available via AT&T.  I now am testing with Galaxy S from HTC.

My take on Android is that it iPhone is designed for ease of use (for most beginner users of smartphone).  Even my wife can use an iPhone and likely she could use my Android but it’s a bit more technical.  Which also translates to flexible?  The android operating system has been the worst in 1 area in particular:  connecting to older versions of Exchange Server.  I assume this is part of the OS and due to its limitations; I had to move to a 20 dollar exchange email app which works great.
In general Android has lots of settings you can configure.  Linux users like a lot of custom settings and advanced editing of the OS tasks.  Android is a lot like Linux.  It seems geared toward the right brained user.

The visual appeal of simplicity is a powerful benefit to any smartphone system.  Windows Phone 7 and iPhone are both simpler for the broadest range of users.  I have been using my Android now for about 2 months and I am not complaining about it except occasionally it has more time out issues.  I like three buttons to move around, I also like how the iPhone navigates.  Ideally, a user should be able to choose how many buttons they want to use regardless of the device buttons.
Not as advanced capabilities in the App Market especially for filtering the list of apps the way Apple does.  Apple at this time seems to have the most mature application for managing “sales” of apps.

The Android feels more like open source in its approach of having “many versions” at once.  There is no easy upgrade path for most devices that run these older versions of the Android OS.  So as devices mature and their life becomes more like the PC of today (I know people that have not upgraded for over 5 years) and many OS updates will be expected to remain relevant.

I sort of like the notifications tab that fills up and sometimes makes me take too many steps to stop it from annoying me.  Android has static icons for applications.  iPhone’s use of the left icon corners and Microsoft’s Windows Phone 7′s panels are a great use of design space.  This is something Android does not have.  It uses icons on the top bar coupled with notification alerts which is less efficient.  The email program on the Android was terrible in my opinion.  Web Links, phone numbers, and formatting are poorly supported in my version.  I switched to Touchdown Pro which does a great job of email for Exchange Server.

Next up…which mobile OS wins this year?

iPhone versus Windows 7 Phone

A few days ago I made the switch from Apple iPhone to Windows 7 Phone.  The day it came out on AT&T.  Not because I disliked my iPhone, but because this new OS will provide a glimpse of how Microsoft will compete for years to come in the explosive Smartphone revolution.  I thought I would write about my experience.  I have not read any instructions on the device (same as iPhone experience).  Both Apple and Windows devices have a significant focus on the younger market (games, social networking, music).  I don’t intent to provide any feedback on the features of either system in most of these areas with the exception of Facebook.  Facebook has a very deep relationship within the Windows Phone experience.  I believe this strategy could drive a major benefit as there are many aspects of the mobile lifestyle that can rely on your definition of “friends” and I expect many revolution applications to drive this forward in the near future.

The screen quality on my HTC is fantastic.

Exchange and Email

First off, I found it slightly easier to connect my iPhone to our implementation of Microsoft Exchange Server.  I had to try changing settings about 4 times before I found the right combination of domain, account, SSL, ect.

Second, I find that lacking cut and paste makes editing email responses more challenging.  Also, I am used to being able to respond to an email (REPLY) and being able to cut some of the content from the email chain.  Windows Phone does not allow this (at least that I can tell).

Synchronizing of email is missing an option that I got used to.  On the Windows Phone, your options are:

• Synch as items arrive
• Various time intervals
• Manually

I like to conserve battery so I am running as “Sync Manually”.  I liked the option on the iPhone that would Synch when I opened the application.  It is just 1 less click.  I find that the Windows Phone interface for mass deletion of emails has an edge over iPhone.

Calendar support is more inclusive of features and capabilities of Microsoft Exchange calendars like recurring meeting support and private meeting support.

Navigation

In general, navigation is straight forward and gestures work in similar fashion to the iPhone.  It took me little time to adapt to it.  I like the dynamic icons on the main screen.  It is an efficient use of space and can provide useful information.  I found the management of icons on both the main page as well as the app navigation a little more complex.  The main area of disappointment for me was the lack of fast navigation through large lists such as with contacts.  The Windows technique to navigate contacts (“People”) is either clicking the search button and start typing, or screen gestures to scroll the list.  With a list of thousands of contacts the process can be improved by taking some concepts like Android and iPhone use such as alphabet drill down which allows the user to click once to drive down to a letter to begin scrolling.

Secondly, coming from another device (Android or iPhone), the sensitivity of screen gestures using your finger is different and takes time to adjust to.  I think a tool that adjusts the sensitivity level much like you find on a PC for the mouse would be a nice enhancement.

In Summary

Will Windows Phone 7 become a dominant player in Smartphone?  Well, based on overall experience, the OS is a contender in the consumer space but its early stages and depth of application availability may be a disadvantage.  The dark horse here for me will be where the future of integration to Facebook will go.  Again, I believe that building deep integration into Facebook could prove to give Microsoft a revolutionary driver. I also believe that this device may dominate the business landscape. Why?  IT infrastructure and management for Microsoft centric corporations are familiar with the tools to support security to back-end systems and integration to CRM, Office, and SharePoint.  In addition, the massive pool of Microsoft developers already employed by these companies are much better positioned to develop business applications than iPhone and Android developers.  The future will be quite a race.

Soft Allocation and the Intelligent License Plate

In my last post, I described a need for a Processing Engine.  This processing engine needs to support very fast operations as well as more loosely defined operations that can only operate more slowly.  By expanding the traditional definition of a license plate, since we now can support more complex collections of goods that are associated with this license plate via the processing engine, we can look at a more complex scenario known as soft allocation.

To explain soft allocation, is to look at a high volume, customer driven (retail) supply chain that must ration its supply.  Large FMC (food mass club) and big box consumer goods challenge the supply chain with concepts such as just in time inventory.  The art of soft allocation is figuring out as quickly as possible the best way to divvy products up based on orders in the system.  Data points such as what supply is on hand and not allocated, who ordered it, how much, and when are typically considered when planning how to distribute these limited products.  Even manual judgements can drive how much the supplier is willing to short a given requester and how quickly more supply is available can define the contents or collection known as virtual license plate.

If an order is treated as a license plate, it can contain quantities and locations of items that are ear marked for that specific order. There is no requirement for a license plate to contain physical inventory and in this case the license plate would be virtual.  It can itself contain a reference to an item and quantity that has been soft allocated based on both system algorithms as well as manual tweaking of who gets what and when.  After this planning occurs, the application does a virtual move of the physical inventory into these “virtual license plates” that are tied to a particular order.  This process clears unallocated inventory found within 1 to many locations (bins, zones, license plates) and moves them to this virtual license plate.  To eliminate the risk of accidently misappropriating inventory to the wrong customer, this virtual license plate can be put onto hold automatically upon creation.

Here is an example of 2 customers both requesting widget A which has limited supply.  Shortages can be common place.

Soft Allocation via Virtual License Plates

Core Elements of the Intelligent License Plate

As mentioned in my last post, the concept around intelligent license plates relies on contextual information (meta data) combined with a scan event to trigger 1 or more workflows that can automate the job of the manufacturer.  In a typical software automation solution, the logic related to all operations exist in the core of the application.  By using this new paradigm, the scan itself can trigger unique business rules outside of the core system.

An ILP (intelligent license plate) system needs a way to manage the existence and archival of unique sequences of digits.  This system should include the ability to insure uniqueness as the use of the system grows.  Note that as the size of digits grow, the size of the barcode needs to also be considered.  An ILP needs the ability to include non-unique barcodes which may be unique for a specific purpose or external party.  An example would be generation of UCC128 data.  Some large customers require specific ranges of data to meet their needs.  The distribution of these numbers must also support client side generation for disconnected operations.  Ranges or unique seeds must be created to avoid duplication of digits.

Secondly, ILP systems need to manage the mapping of the barcode + metadata to a set of workflows or states within a state machine.  The workflow management should allow for synchronous and asynchronous execution.  Asynchronous would be a good candidate for secondary reporting data that drives a data warehouse or reporting cube.  In scenarios where the mapping results in a sparse matrix (many possible inputs) and a workflow is known, ranges of values must be easy to define.  In scenarios where a workflow is not known, a special workflow termed “universal state” should be in place as a sort of “catch all” for loosely defined or highly dynamic business rules.

ILP Processing Engine consumes the inputs and processes the state machines and workflows.  It is tightly coupled with the ability to query various areas of the core system as well as generate external calls (service end points).  It must transact inventory operations (change, move, transform), generate new ILPs, produce audit trails, and generate user messages.  The processing engine is responsible to providing a standardized approach to all these areas.  To deal with a variety of needs, the processing engine must support fast operations, and slower loosely coupled operations that must traverse large sparse matrix of states.

Intelligent License Plates and State Machines

Intelligent License Plates Part 2

We start with the concept of a barcode.  A barcode is almost magical if you think about it.  A barcode can mean almost anything…literally.  A barcode does not need to be long and packed with digits to make this statement true.  For those familiar with database technology, a barcode can be compared to a primary key or index into a set of information.  The larger the primary key, the larger the collection or set of referenced objects.  The interesting idea around intelligent license plates is that the primary key can be comprised of both contextual information as well as the “Scanned” barcode.  For example, lets say I scan a barcode “12345678″.  I could add contextual information to this about who I am, what time it is, and what operation I am working on.  Therefore, the primary key into a collection of objects, or in this example, workflows, could be represented as “TSMITH+PUTAWAY+06/25/2010+12345678″ or more realistically the user name, operation, and date would be translated into integers and the primary key would become something like “0010+034+06252010+12345678″ or “00100340625201012345678″.  By extrapolating the context in which the scan occurred, we can now apply an additional layer of insight into what steps need to be taken or to put it simply, what workflow(s) need to be triggered.  There are now 5 possible primary keys that we can use to allow the system to intelligently find automation steps.  1 for user, operation, date, barcode, and the combination of all 4 primary keys.  The large primary key allows for storing data in a very sparce matrix.  (A topic for future discussion…)

Now that we have our context established, we can couple this with historical information (previous state information) to input into a technology known as state machines. 

If you are not familiar with State Machines, it is a concept that originated in computational theory and mathematical models for biosciences and mathematics.  It was then adopted in the 1960’s within computer science for language parsing and compilers.  This pioneering work is largely attributed to the efforts of George H. Mealy and Edward F. Moore.

The basic operation of a Finite State Machine system is this: as the system is in one of the defined states (represented by our primary key) at any instant of time, it will react to specified (external) inputs or (internal) conditions with specified actions (defined in workflows), and transition to another defined state, or remain in its current state, depending on the design.

These workflows define what we want to occur when the scanning happens.  For example a workflow may have rules such as, if a supervisor scans a license plate that contains expired products during a Quality Assurance operation, then put the license plate on hold and send an alert to purchasing and management, that there is expired product that has been in the warehouse for X amount of days.

US Manufacturing Needs a Boost of Innovation.

highroadnetwork.org/2009/11/10/67

Intelligent License Plates

I am publishing a technique for process automation, in hopes for cost savings to manufacturers in America and abroad.  I believe that by implementing Intelligent License Plates, supply chain businesses will benefit and hopefully save on operating costs.  Whether we want to believe it or not, a lot of people will buy the cheapest products that suit their taste.  Period.  We believe our technique will allow companies that run poorly in meeting a customer’s needs to utilize Intelligent License Plates.

In barcode scanning for warehousing and manufacturing, one size rarely fits all.  A workflow for a process that works for one organization may be cumbersome or viewed as overkill for another.  The variables involved in optimizing an organization’s processes are often too complex to assume one workflow fits all.  Even within one organization, the optimal workflow of automation required may vary greatly by time, date, sales region, customer, user, products, or locations. 

Variables such as workforce enablement (trusted or not trusted operations), volume, product variability, and probability of outcomes during an operation are all important factors in tailoring a solution that works right. 

A standard license plate is a unique tag or label that can be identified by a scan or RFID read.  It represents a multitude of information about a collection of items associated with this license plate.  Information such as items, lot numbers, serial numbers, and unit of measure (UOM) are examples of a collection of items.

The concept of an intelligent license plate takes this concept and goes further by incorporating the state of the license plate in the process (some examples are new, just received, just put away, just produced, just picked, just shipped) and combines it with the context of a scan which includes, what part of the application the scan occurs and who did the scan.  This optimizes labor and automation in a unique way.  This technique optimizes operational processes around the most likely outcomes by removing the burden of focusing on exceptions.  Exceptions, while a reality in day to day work, should not hinder the speed of the most probable outcomes.

Intelligent license plating creates a workflow plan for a class or set of behaviors for license plates that automates certain tasks based on what state the license plate is in, what operation is being used, and who is doing the operation.  Before a product is manufactured, an operator can create a workflow that defines what should happen when someone in manufacturing scans a new license plate associated with this specific workflow.

To better explain intelligent license plates, an example seems appropriate.

A food manufacturing company produces pallets of finished goods in 2 different manufacturing locations.  In one location, they wish to produce products that have very low variance from pallet to pallet.  Because of this low variance in the finished goods and raw materials, they decide to automatically create the finished good and back flush the raw materials.  In addition, they always move these pallets into a pre determined location for order fulfillment in a specific zone.

In a second location, they produce products that have high variance from pallet to pallet and have a need to do a quality assurance hold until results are tested.  In this location, pallets can be placed into a variety of zones.

If a supervisor scans a license plate while a license plate is new and it exists in the manufacturing location, they would like to automatically assume that the supervisor wants to review the license plate and potentially change its hold status.

To automate both of these processes, this company would create 2 license plate workflows.