Putting the "e" in Customer Content: An Intel Case Study
How does a company stay agile in Internet time? At Intel, many of the answers were gleaned from the electronic content experience.
Everyone is familiar with the continuing saga of the little "e" that is changing the shape of business. Back in the mid-'90s when the Internet gained a stronghold as the next communications medium, every company raced to provide at least a URL for their customers to find them on the Internet. Then companies began to strategize on where to go with this new "e" thing. As a first step, many IT and IS muscles started flexing around programs that put the "e" in front of commerce. By moving order management to the Internet, companies can provide better Customer Relationship Management simply by making it easier to do business with them. Intel Corporation moved virtually 100% of its customer revenue to the Internet over two years, starting in July 1998.
What else can IT and IS do with the vehemently powerful little "e"? At Intel, the next step was to move thousands of specific customer documents and content to the Internet. This was no small job. Intel has thousands of worldwide customers that need hundreds of thousands of various documents ranging from highly confidential microprocessor specification drawings tailored to particular customer products, to channel data that pertains to sub-segments of resellers around the world. Intel and its customers were used to a slow manual process of sending and receiving confidential documents. In an industry where time to market is a make it or break it factor, there had to be a better way.
To eliminate the time and cost drain of the old-fashioned content delivery method, the corporation focused on creating a single electronic content engine that could simplify and streamline customer content and document delivery. Getting to one electronic content and document delivery system using the Internet was hard.
Here's how the corporation's e-Business information systems team created an electronic environment that provided 100% of customers 100% of the documents and content they need. From business process change to IS re-engineering, the resulting challenges and learnings are invaluable to any "e" initiative.
Open the information flow for the customer.
o Spend up-front learning time with your customer.
o Let the customer manage their experience.
Facilitate business process change.
o Anticipate significant process refinements.
o Architect for choice.
Utilize robust core services.
o Share core services amongst customer groups.
o Use Commercial Off the Shelf solutions (COTS).
Stay agile in Internet time.
o Prepare for the long run but operate in fast release cycles.
o Modify best-of-breed applications.
Environment Leading to e-Business
When Intel launched its e-Business website in mid-1998, the corporation was undergoing some major changes. Business was shifting around the globe. Intel's bread and butter product lines of microprocessors and embedded silicon chips were rapidly expanding into communications and networking silicon. The percentage of revenue from the Americas was shrinking and revenue in other geographies was growing. The worldwide channel of distributors, resellers, and system integrators was rapidly growing in percentage of Intel business and processor market segment share. These significant changes created an environment where Intel could aggressively embrace the Internet as a communications tool to help expand its reach and depth into existing accounts and new channels.e-Business Value Proposition
In 1998, Intel's Customer e-Business Mission was "Add significant value to Intel customers, brands and products by using the Internet." The tenets of the value proposition were:
o Improve Customer Responsiveness - Breadth and Depth in Customer Accounts
o Optimize Time to Value - Automate Content Delivery
o Expand Customer Reach - Global Consistency
o Increase efficiency and productivity - 24 x 7 Customer Support
The Customer
Intel has two types of customers: direct and indirect. Direct customers buy products and services directly from the corporation and have direct support from its field sales organization. These are mainly Original Equipment Manufacturers (OEMs) and distributors.
OEMs develop their products around Intel's platform roadmaps. By providing OEMs with products and product information well in advance of when a product becomes public, most customers can ship new products on or about the same day. Any information sent prior to the product release date is confidential and only supplied to the customer if a valid Corporate Non-Disclosure Agreement (CNDA) is recorded. A Confidential Information Transmittal Requests (CITR) also must be signed and returned for each document before any actual documents are released to the salesperson. Before deploying the e-Business site that facilitates electronic delivery, all CNDAs and CITRs had to be hand delivered for each document to each individual, company, and site.
Indirect customers vary within a wide range of size and scope. There are roughly 100,000 various types of channel customers; resellers, systems integrators, independent software vendors, and solutions providers that Intel works with in a channel sales capacity. The Internet is used for product and technical support because customers typically buy products and services from Intel direct customers. The end user, or consumer, is considered a subset of the indirect customer group.
Specific content and messages are associated with each particular customer group. Intel views the direct and indirect channel customers in a business to business, or B2B, relationship, whereas the consumer and end user customer experience is categorized as a business to consumer, or B2C, relationship.
The Sales and Marketing Funnel
From initial awareness to post sales support, each phase of the sales cycle is supported via the Internet. The Internet serves to distribute content by allowing Intel to segment and target content to specific customer sites. For Intel's OEM customers this often includes focused sales and engineering support for designing the Intel product into the customer's product line. These embedded products need customized technical support documents on configuration options, upgrades, errata, etc., for the changes that may occur in that product life cycle. Public content created for indirect customers and end users is very different than the more sensitive technical and business documents needed by direct customers.
Beginnings of e-Business
Intel's foray into Internet Customer Relationship Management started in late 1994 with the FDIV crisis. The Floating-Point Divide Error (FDIV) was a hardware flaw in the Intel® Pentium processor. After a customer discovered the flaw, a media frenzy around what Intel would do about it quickly ensued. The corporation decided to replace all faulty units. The corporation used the Internet site it already had in place with technical customers to communicate the replacement policy to the public. Through this experience, Intel learned the Internet was an amazing communications vehicle for getting messages and information out to a broad audience. Intel has since made known information on errata (hardware "bugs" or flaws like FDIV) publicly available and used the Internet to facilitate that goal. The company thus began managing customer relationships via the Internet.
End users and consumers typically go to Intel's corporate site, intel.com, for information on Intel products, technologies, management and press releases, stock and financial information. Indirect channel customers go to channel.intel.com, which has comprehensive local language support and geographically specific channel programs and products. Developer.intel.com is targeted to engineer customers, which need publicly available technical documents to help "design in" Intel products into their products. Support.intel.com is a product support site that allows any type of Intel customer to obtain post sales product support over the Internet.
In January 1998 the company started working on the first true e-Business customer offering with a website for Direct Customers. After six months of development time, in July 1998, Intel launched the Intel e-Business site which included content and document management, along with the first iteration of e-commerce applications. After the first 15 days of operation, Intel booked $1 billion in customer orders from about 20 customers.
Content Management - A Brief Overview
Currently, most marketing and consumer oriented information is offered in HTML pages referred to as "content." For intel.com and other Internet marketing sites, the company uses a content management system that supports and tracks over 300,000 documents and pages of Web content. A centralized content management system lets Intel store and manage most of the content in one place. Each website has its own directory with different access rights. Roughly 350 content authors at Intel save their text documents in the data directory. They specify post time, and the document is either translated into a Web page (HTML) or posted as is, on the appropriate site by a Web author. Non-U.S.-based content authors do not typically use the central system due to global issues of limited infrastructure and slow connectivity. Instead, local agencies send the data to authors in the U.S.
Document Management - First Revision of Information Desk
The first iteration of confidential document delivery was what Intel called Information Desk. In the initial version, when a document request occurred, the system would actually sort through all 300,000 available documents and content pages within the intel.com family of websites. The initial goal of the first iteration of Information Desk was to make it a document navigation tool. 90% of direct customer confidential information, however, is created in regular office applications, not HTML pages.
Confidential documents consist of both Intel and customer documents so Intel had to be particularly cautious with devising a document distribution system. The system needed a high level of access control. The e-Business team discussed the CITR agreements with legal and information security groups and devised a policy to manage and use electronic CITRs. The system kept track of who saw what, and prompted customers to accept the eCITR the first time they viewed a document on the e-Business site. The process is similar to a software license agreement.
To take full advantage of the speed and efficiency of the Internet, and to offer customers segmented services, software engineers decided to create a separate document database for confidential documents serving direct customers. Engineers integrated confidential document distribution needs with Intel's backend document management system and added a targeting feature so Intel could indicate which company received which document. One content engine served all customer segments.
Approval and Authoring Systems
Simultaneously, Intel developed an Intranet document authoring and approval system as well as an access control application to allow the documents to flow from inside Intel to the customer. Field Division Business Link (FDBL) is an Intranet site and application that allows divisional and geography marketing and technical authors to send confidential documents to field sales. They in turn can authorize the information is sent to customers over the Intel e-Business Site via Information Desk.
Access Manager is a similar Intranet application built on the FDBL platform that allows Intel employees to control the passwords and level of entitlement assigned to customers. Access Manager and FDBL were both designed as a centralized streamlined system. One underlying document management system links Information Desk, FDBL, and Access Manager applications together.
By the end of the first three months of operation, it was apparent that the design encountered two major challenges. One result was that both internal and external customers were suffering from information overload. The several hundred authors developing the confidential documents did not know exactly where the document went, or who would ultimately be reading it because they passed it to a database rather than handing it to a sales person or customer as they used to do. So they sent the documents to as many sales people and account teams as they could to make sure their customers were covered. This bogged down the system and made searches painfully slow.
Even worse, the sales people were not approving the plethora of documents, thus slowing down the process even more. The field account teams knew exactly where the content belonged but did not have the bandwidth to sift through, read, and approve the high volume coming from the divisions. The problem was further compounded as sales people dialed into the system from on the road via low speed modem connections.
Automated Approval is the Answer
The team quickly became aware of a need to balance two competing approval philosophies. One, let everything go out to the customer and potentially violate security policies. Two, have the field approve every document before it goes out to the customer, and have sales spend their time approving content rather than selling. The approval answer was a balance somewhere in between the two.
In the latest subscription service, responsibilities are shared amongst various business and technical groups. The sales people create the rules and business process for automatically sending documents to their customers. The technical team then has the challenge of taking those business process rules and enforcing them into the approval application. A Product Manager consolidates and represents all the business needs, and works with the engineering team. An e-Business team identifies the right rules, and implements a refined automatic document approval process. The result is a simplified approval process for the field that maintains the integrity of confidential documents.
Information Desk - Initial Results
External customers really liked getting the documents online and not having to wait for their Intel sales people to deliver them. The Internet saved them time and allowed them to design in Intel products faster. Navigation and ease of use, however, were not very slick, and coupled with long waits for information to flow through the network, the external customer experience was satisfactory but not delightful.
The internal customer was not as satisfied. Sales people were used to a process of getting the customer related information from divisions that had evolved over the years. They were in the habit of either calling someone, using email, or going to different divisional Intranet sites to get the information. Even though FDBL was one central place to go for standardized content, the delivery was too slow and bogged down by the sheer amount of documents in the system.
Based on worldwide customer feedback many enhancements were made to the Information Desk application. A second cleaner version released in Q399 included a "What's New" section and a customizable category feature in the Information Desk area. This allowed customers to see the latest three to five documents. It also let them choose to view only the content categories that mattered to them. This significantly quickened the download time. A "What's New" section was placed directly on the e-Business Site home page, to allow customers to find what they were looking for much faster, with fewer clicks.
Initial Misses
In hindsight, the e-Business team identified three gaps in the customer experience. One, the overall ability of the customer to get to the information was not clean and simple. Two, architectural issues with the backend made the system too slow. Three, the business process did not easily sustain new document creation and flow. They knew that for customer's to truly embrace the Information Desk and FDBL applications, they had to significantly improve the speed of access and offer more customer oriented functionality.
In Q1 2000, Intel released Information Desk II, a true rewrite of the application. This release moved the content application squarely into the "enablement" phase of e-Business development. Rather than merely automating information delivery, both internal and external customers were empowered to quickly decide what to read and when to read it.
Information Desk II - The Technical Solutions
To afford the needed functionality improvements, Intel cleaned up and revamped the entire system. The team identified five areas of improvement that would enable vast speed increases to internal and external customers.
First, Speed and Ease of Use Enhancements for Information Desk
The first goal was to make the pages easier to navigate and the documents easier to retrieve. For document categorization, the team increased from a two level to an "N" level categorization. Information was further classified so users had an easier time finding the documents. Based on customer feedback, Intel rearranged the page layout the way customers wanted to receive information, instead of how Intel engineers thought they did. A pop-up navigation feature was included so customers could review the document without opening it. Finally, the launch of a document search tool filled a critical need.
Second, Significantly Improve Usability
A business and technical SWAT team was formed to assess the best and fastest ease of use changes that could be made to FDBL. The team came back with HTML home page rewrites that provided a much better user experience. The new look and feel is much easier to navigate and use. The team learned a "keep it simple" rule. Exquisite technical solutions are sometimes not the best customer-oriented approach. Plus the perfect technical solution tends to take longer than Internet time allows. The business team worked with the technical team to create a better customer oriented solution, though while not a technical marvel, proved to be exactly what the customer wanted and needed. This was the first of many ongoing usability upgrades.
Third, Break Out B2B & B2C Content
The e-Business team separated the infrastructure of the 10,000 B2B confidential documents from the 300,000 documents and HTML pages in the mostly B2C content repository. By segmenting by audience, customers received the information they needed much faster and easier. With a separate B2B document repository, a complete search would be at worst looking through the meta data tags of 10,000 documents. There was no compelling business reason to keep the information systems together if pieces could be broken apart to offer better functionality. The team could perform changes faster without having to wait for the consumer, B2C, side of the intel.com offerings to be ready. Finally, by moving to a dedicated new B2B infrastructure the team had a chance to create a scalable flexible long-term document management solution.
Fourth, Fix It at the Root - The Database Engine
Another reason for the slowness of the overall system was the way Information Desk and Access Manager's used the database structure in which it was built. Many simultaneous queries caused data to lock in the table structure, which caused system crashes. The team chose to install a completely new database engine, based on technology the team uses to build many of its major web applications. The new system was different, however, then the one used to create the existing electronic content and document management systems. The database schema and data model were preserved intact, just the database engine was changed.
Fifth, Make It a One-Stop
The team had to link together the information on external customer, and internal field, geography and division author access rights, including levels of confidentiality, as well as parameter definitions, into one massive yet exquisitely efficient underlying content management system. They needed to get to a one-stop shop for files and metadata. That meant merging the application architectures of the customer databases: Information Desk, FDBL, and Access Manager. This was a rather large undertaking mainly for the business process analysts. They pared down the business rules and approval steps the internal customer experience entailed until the resulting tool was streamlined and efficient.
Technically, they achieved a one-stop shop look and feel by doing three things. One, the different teams working on FDBL, Information Desk, and Access Manager used the same data modeler. This person negotiated between the Database Analyst and application development teams for all three projects. Two, the three teams met consistently to talk about and agree on the definition of fields and usage of the shared data. All data was looked at, set, and consumed in the same ways by all three systems. Third, new database links allowed the teams to use a master document and metadata repository. The data may look separate to customers but really is held in the same repository, which keeps the system centralized and manageable.
Key Learnings from Information Desk II
Merging three systems while upgrading and adding significant enhancements to them, and disconnecting from a larger content pipeline was a hard order to fill. As a matter of fact, the Information Desk II release nearly broke the team. They ran into a number of issues due to the fact that everything in the e-Business Site application architecture was changing except the e-commerce applications. This was the biggest change for the team to make and the margin of human error was at it's most vulnerable. Following are the top learnings as seen from their eyes:
o Automate the Build - Between the e-Business Information Systems and the Information Technology team, they manually performed thousands of tasks to set up the servers for the various upgraded applications. The team made a rule to have a well thought out documented set of build processes for every new release, regardless of timeframe or size. In conjunction, IT and IS team up to automate the process for faster server builds. This minimizes development time and margin of error.
o Test Harness - Every application layer was dependent on another, yet various owners were in charge of quality control of their application. The team implemented a test harness that calls in any dependencies and can quickly ascertain whether an application or interface is wrong. Thus, the margin of human error in detecting incompatibilities is minimized.
o Organizational Evolutions to e-Business - Three separate technical teams as well as three different business teams groups co-owned the three applications that were being upgraded and rewritten. Between the six groups, there was no Master Project Manager defined. Project deliverable goals such as deadlines were not in synch. Many ergs were spent figuring out how and when the various application changes and launches fit in with other group's goals. By organizing a central e-Business group responsible for all business and technical decisions, the frantic pace of Internet change and subsequent development can be managed.
o No Mass Releases - With hindsight, the team would never have done so much at once. They upgraded the main database software running the applications while merging pieces of the architecture together, and rewriting the business processes. Plus, they broke out a complicated subset of the application. Lesson learned: Don't intersect all the releases and changes into one. The margin of error increases with every radical application architecture move. With several releases happening at the same time it is very difficult to plan and protect against unforeseen delays and errors. Indeed, the Information Desk II release was four months late.
o Faster Cycle Spin - Engineers are used to 12 to 18-month release cycles. Herein lies a key challenge to IS and IT organizations today. A new challenge is how to put processes and organizational rigidity in place for faster development. At the same time, the customer is demanding more and more easy-to-use Internet savvy applications. The Content Management team at Intel has learned to operate in a 3 to 6-month cycle, from initial idea to release.
As Sandra Morris, Vice President of Intel's e-Business Group likes to tell her employees, "We are driving an 18-wheel semi down the highway at 90 miles an hour and it's our job to change the tires while keeping the course and speed of the company and customer expectations." Having the right people in place to steer the truck, navigate the next turn, change the tire, keep the tank full, and eye the finish line, is the challenge of any e-Business team.
Electronic Design in Services Go Beyond Document Management
To adequately support the Design Engineer customer Intel had to do more than automate confidential document distribution. Design Engineers need support for issue management and post release product support. Other areas of electronic design in support at Intel are the Sales Knowledge Database (SKB) and the Electronic Design Kit (EDK).
The SKB offers all the relevant technical documentation the field sales teams need in one package. Unlike FDBL, which offers the documents one by one, a SKB lets the field download the entire package. Customer Quarterly Product Updates are an example of this bundled content. SKBs are housed on simple HTML pages that link into the larger FDBL infrastructure. This is an example of a customer centric device that let's the field get the information they need in the most efficient way. This saves the field from sifting and searching through the larger pool of available documents.
EDK is another Internet application Intel uses to distribute necessary design in tools and documents to indirect customers. Like Information Desk, EDKs enable customers to build designs over the Internet. The difference is, an EDK is an aggregated set of integrated tools that bundle all of the technical documents and resources needed, and offer this package to Intel indirect customer design engineers in one click. They are currently available for non-managed designs, on developer.intel.com. Non managed designs mean there is no Intel technical support contact working on the design with the customer. It also means the designs are public information (i.e., there are no eCITRs or confidential documents included). The EDK application is linked to the same content repository the rest of intel.com uses. EDKs are delivered only when the customer needs them. Basically, as the SKB is to the field sales team, the eDK is to the Indirect design engineer customer.
The EDKs were launched in mid-2000, with an initial 20 EDKs focusing on communication embedded silicon products. EDKs are linked off of Information Desk and reside on FDBL. A handful of beta customers are piloting the EDKs on Intel's e-Business Site for direct customers.
The EDK application also offers more interactive tools including validation files that run against customer design. These let the customer upload their files to the Intel site of their preference to run it against Intel's specification standards. This offers immediate verification instead of using the traditional "diskette in the mail" approach. Another exciting feature on this platform is the Dynamic EDK. By using metadata tags, relevant components of the design kits will automatically be created dynamically upon the creation of the platform. The EDKs are then automatically created as the customer requests. Ensuring the customer gets the information needed in the fastest most efficient way possible is what customer relationship management is all about.
EDKs are a great example of content delivery that works for the customer. They represent breakthroughs by offering more targeted content to the right customers. In the long run, however, they are still an interim step to the true personalization path. Intel is not reaching a complete personalization level until only the exact data needed is extracted from the various documents and inserted into customer's preferred receiving method.
The Channel Emerges - Indirect Customer Content Management
Intel's indirect customers consist of hundreds of thousands of resellers, system integrators, networking and e-Business solution providers, as well as design engineers at indirect OEMs and Distributors. Intel channel customers sell system products, like networking switches and routers, that are finished goods needing to be installed and configured into end user environments. The indirect customer base has been steadily growing in breadth and depth of Intel product penetration, market segment share, and strategic importance to the computer industry and Internet economy. As such, the number of Intel boxed and branded system products and corresponding reseller programs have substantially grown.
In addition, the channel customer is demanding better and better Internet tools from their key vendors. The corporation had to ramp up customer relationship management via the Internet as the most efficient way to support this increasingly important customer base. Channel.intel.com and developer.intel.com sites are the mainstays of content for the indirect customer. For over five years these sites have delivered the specific program and product-related content and documents customers need to create, market, sell, and support Intel products.
A new and distinct area of content distribution developed for indirect channel customers. As the indirect customer base grew, there was an increasing need for more content of a semi private though not confidential nature. For instance, Intel keeps track of reseller rebates online. This is very sensitive customer data but the content is not an Intel confidential document it is part of a database that has extremely limited access. The e-Business team, therefore, turned its attention on developing a robust channel content management systems.
Common Login Provides Customer Centric Synergy
The first step in creating a channel focused content strategy was to create a common login. In mid-1999, Intel created a program that allowed all the customers to log in to one place, even though they belonged to several separate programs. For instance, a reseller could be an authorized solution provider for Intel networking, processors, and e-Business products. Before they had to login to Channel.intel.com with three separate passwords and get directed to three separate areas.
With a common login, the team moved to a One Intel feel. Customers are now differentiated by who they are not what programs they belong to. The common login has since turned into a core service that allows Intel to bring communities together through a common experience. This in turn is the basis for a much better online branding strategy. Common login is also an example of vendor centric customer relationship management evolving toward a customer centric mode.
The indirect applications team built the common login off the already existing core service developed by the direct customer applications team. They first examined the functionality, policies, and dependencies of the authentication core service. Then they separated out the key core technologies, and broadened the access pieces of the application to provide for the indirect channel customers. The result is a flexible core service that provides a way for the resellers to create their own usernames and passwords. Login pages can be personalized in a centralized mode, or local geographies can customize them as well.
Semi-Private Content Management
After building the common login for indirect customers, Intel could redesign its content architecture to make personalized content available to resellers, as they needed it.
The team decided to create a third content infrastructure to support the indirect customer's increased need for personalized semi private content. They expanded the current content management platform with a new product that specialized in content creation and delivery, rather than in document distribution. The new semi private infrastructure allowed the corporation to support the channel's local language as well as connectivity needs. By using an application that focused on content workflow rather than document management, the team quickened the speed of information to the customer and provide the content in the best possible view for the customer.
Content Delivery for the Customer
In summary, the corporation's content management strategy evolved to remain focused on the ever-changing needs of its customers. The corporation now has three separate content engines: one for consumers, one for direct customers, and now one for the semi-private content needs of channel customers. The systems do share certain pieces of the same core services, like login and authentication. They will continue to build off of each other's learnings moving forward. For instance, based on the complex problems solved in the local language support work done by the indirect applications team, Intel hopes to create a local language core service that can be used by all Intel websites.
Figure 1 - Intel's Content Management Platforms
Overall Results
By the end of 2000, after almost three years of intense work around using the Internet to increase customer satisfaction and enable a closer and better customer relationship, the team reassessed its efforts. The technical team moved from a basic IS group to a leading edge e-Business commando team. The cultural change at the corporation was palpable as well. Although there were groups still slow or reluctant to embrace the changes associated with new Internet based tools, it was clear to all that the Internet was the future of customer relationship management. Finally, to fully analyze the impact of the Internet on Intel's CRM strategy, a small team of e-Business analysts rolled up the return on investment data and customer feedback on the "e" content and document management efforts. Here are highlights of what the initiative has accomplished so far:
Consistent Reach to All Customers
One of the first imperatives was to level the time to information between various customers and between customer locations, so that they received important information at the same time. Using traditional hand delivery of confidential documents created customer disparity between time to information based on proximity and physical location. For instance, a sales person in San Jose, CA could get a confidential document to his account's headquarter location the same day. His counterpart in Japan may need to wait a week for interoffice mail and then travel eight hours to deliver the same document to the customers' Japanese regional offices. In addition, there are salespeople who focus on one large customer, while others focus on several smaller accounts. In the channel space, Intel could not possibly or cost effectively reach 100,000 resellers (indirect customers) with direct sales efforts, plus the resellers did not need that level of account management.
This "tyranny of distance" was shaping the customer relationship in a way that made it impossible for Intel to manage. By using the Internet as a real-time content delivery vehicle, the corporation automated and standardized processes around information delivery to let the right customers get the same data at the same time. Now a direct customer receives Intel confidential information at the same time at its headquarters and remote locations. Channel customers have equal access to their sensitive program information. Large sized accounts can get the same attention as smaller accounts by using Information Desk to receive the latest Intel information. The Internet is now a primary vehicle used to increase customer time to value and time to information.
Extending the Reach for Indirect Customers
Using the Internet to better manage reseller programs allowed the corporation to reach more customers. By standardizing the sales and marketing messages on the Internet, the sales teams had a better way to consistently manage the demand of Intel technology products. For instance, Intel created a program for the Boxed Processors product line that provided end users certificates of authenticity, registration, and special offers, only through authorized resellers. This served as an incentive for resellers to sell the boxed products instead of other unapproved processors. The program was managed exclusively on the Internet because it was the only way Intel could offer consistent global support for over 75,000 resellers. Intel used the Internet to expand its reach in the channel as well as offer a more consistent quality product.
Increase the Time to Information
Information Desk has significantly reduced the time it takes Intel direct customers to receive confidential documentation.
Before Intel's e-Business site it typically took between two and three weeks for a document to travel from a Division or Geography content author to the Field sales person, and then to the customer. After the first release of Intel's e-Business site that was squeezed down to an average of three days. After the Information Desk II release that time went down to 30 minutes.
Three out of four of Intel's direct customer engineers claim that the new electronic method of receiving the confidential documents has shaved off at least a week in their overall product development time. This figure equates to real dollars saved in today's fast-paced Internet Economy. The corporation's electronic content and document delivery system also serves to increase the customer's visibility in an increasingly volatile and uncertain economic climate.
Overall Key Learnings
The technical team had some important key learnings they will stick to moving forward.
Use Best-of-Breed Commercial Off-the-Shelf Software
The team decided that going forward, all new core services and platforms would be based on best-of-breed commercial off-the-shelf (COTS) applications. This maxim applies to both existing application development as well as new decisions. They did this for three reasons.
One, faster time to solution means better customer responsiveness. Instead of modifying or customizing applications, choose COTS solutions that deliver the closest customer fit. This allows Intel to offer a greater breadth of solutions faster. Plus, it makes it easier to integrate COTS applications as e-Business standards become more prevalent.
Two, cost and developer savings. The less an application is customized and modified, the easier it is to maintain and upgrade. The team realized that on average 50% of development time was spent modifying existing already customized applications when making enhancements or upgrades to the system. Also, every modification had a domino effect on the rest of the system. This scenario brought down IS productivity as well as the ability to maintain robust core services. In the long run, the modified application is much harder and more costly to maintain, while the COTS solution is much easier to upgrade.
Third, safer and easier deployment model. When various applications perform together to create one e-Business solution, upgrading can be dicey. However, by insisting on standard interfaces, COTS solutions extend the flexibility and availability benefits of server scale out to software, simplifying integration, testing and maintenance.
Plan For the Long Run, but Release in Fast Cycles
Soon one of the only real competitive advantages left in the new Internet Economy will be how fast a company can respond to and change their business and customer experience. Today if a company can not get a new feature implemented into their customer site within three to four months, they might as well kiss that customer goodbye. The corporation discovered that the following four areas are the key factors in determining how agile and nimble they could be in Internet time.
o A flexible scalable architecture at the base
o Core platform application services
o Best of breed Internet applications
o Exquisite business process change
Moving Forward
Moving forward, the customer e-Business team is focusing on one thing over everything else, the Customer Experience. How can they continue to deliver the information their broad customer base needs in a faster and more efficient, personalized way? Here are some initiatives on the immediate horizon:
Figure 2 - Faster E-Business Release and Innovation Cycle
New Personalization Platform
Intel is pushing the envelope in offering customers more and more personalized Internet experiences while maximizing the new reach the Internet offers. The latest example of this is both a marketing program and a new content delivery platform. A new Internet based vehicle will provide marketing and support for a new customer group. Intel hopes to reach about 10,000 new customers via this new Internet channel. That is significantly more than the current cadre the sales people can support.
From an infrastructure and platform strategy, the team decided to further expand content management delivery to support this new initiative. The implementation of a personalization platform allows the technical team to develop content delivery solutions that more easily enable the delivery of the right content to the right people. Implementing a new personalization delivery platform will double the team's development time. If the team force fitted the other two platforms into this personalized space by customizing them, they could be done faster but not as cleanly.
The new platform is built for personalization. For instance, the time to write code to change the rules for personalization is significantly reduced with a best of breed personalization application, rather than a customized broad management platform. It used to take months to get one personalization rule through the system. The new personalization platform is much more responsive. It provides for instant changes in the rules engine via a GUI interface.
Flow Content to the Customer
The team is focusing on how to open up the pipe and flow content out to customers with a better subscription service. They can get 80% of documents to flow automatically by entitling customers to appropriate groups of content instead of entitling individual document to individual customers.
The Right Content at the Right Time
The team has a new mantra "Give the customer only the right info at the right time". Only the right content may mean delivering 10% of "A" document, 20% of "B" document, 5% of "C" document, and delivering it in the customer's preferred mode. The goal is to automatically send customers every piece of relevant information on the product and technology the need, but not the 50% to 80% more information they do not need.
Critical Technology Advancements
The appropriate advancement in technology is critical in reaching these goals. Standards in metadata tags, content formatting, and parameters based on key agreed upon data terms need to be created. They must be broadly supported across the channel. Anytime anywhere rules of customer experience need to further evolve. Many of these next steps are based on technology.
The e-Business team is working with many application developers to continue pushing the envelope on creating customer centric solutions similar to the ones it has already started creating.
Conclusion
The real overarching question for e-Business architects today is, how does a company stay agile in Internet time? At Intel many of the answers were gleaned from the electronic content experience. First and foremost, business process change must happen. After the first round of changes, anticipate significant process refinements to continue.
Second, Architect for choice. Prepare for the long run but operate in fast release cycles. This is the only way to stay ahead of the little "e" that is changing the shape of business. Utilize robust core services and share them amongst customer groups. By all means, use Commercial Off the Shelf solutions (COTS) and Best of Breed Applications wherever possible, with as little modification as possible. This will save countless hours and dollars in the long run.
Overall, the electronic content manage-ment system experience has served the cor-poration well. The key learnings will be used in other e-Business projects at Intel. The team learned the perils of overengineering and the joys of simple and easy to use cus-tomer enhancements that enable customers to manage their own experience. Thus, an electronic information flow for the customer, not just to the customer, does create better customer service.
