Like many years before it, the year 2000 was meant to be the year that wireless computing really took off. Instead, 2000 saw the bursting of the dotcom bubble and the much criticised introduction of WAP in Europe. For investment banks wanting to introduce wireless services, these events raise the question: will wireless technology ever provide truly valuable services to them and to their customers? This article explores the business and technical issues surrounding the deployment of wireless services and proposes that, in order to succeed, investment banks must be able to articulate both a business and technical vision. Neither alone will suffice.
I Have A Dream
It is now the received wisdom that wireless services cannot live alone. You build a wireless service to complement services delivered over other channels, such as branches, direct sales, telephone, the Internet and even interactive TV. Most believe that the services delivered over the wireless channel should be a carefully selected subset of those available on other channels. You may offer news headlines on mobile phones, with the full news stories only available over the Internet.
An example of the kind of multi-channel strategy being adopted by the investment banks is provided by Al-Noor Ramji, Global Chief information Officer at Dresdner Kleinwort Wasserstein (DrKW). In late 2000 DrKW announced their "Personal Pocket Portal", allowing their clients to access the full range of its investment banking services from wireless devices. Ramji explained, "The Personal Pocket Portal is now being introduced to clients as an extension of our existing web services. It means our clients will be able to access DrKW's Globeweb Equity Research, FX Order Entry, Indications of Interest, Equity Order Entry and Order Book Browsing, and Trade Settlement and account back office information via their wireless device of choice, such as a mobile phone or PDA. These services will be available to clients via the best available secure wireless technology in their local geography, for example, i-Mode in Japan and GSM or GPRS in Europe."
Ramji explained that the services offered by DrKW "ultimately reduce our clients' cost of doing business". How is this achieved? By opening up the innards of DrKW to its customers, they remove the need for customers to replicate the same functions. A good example is DrKW's 'ops/tracker' online settlements tracking system, which was named best site for cash equity settlement in Euromoney.com's e-commerce awards. Ops/tracker provides a view for both internal users and clients of the status of trade settlements for which DrKW has acted as broker. So clients can potentially reduce the amount of settlement tracking activity carried out in-house; instead they can rely on DrKW tracking and monitoring on their behalf, alerting them whenever action is needed. Where does this lead? At the extreme, investment banks could extend their roles to provision of virtual back offices for their clients, leaving clients to focus on what they do best: managing the assets of their customers.
This works for the front office too. Providing full access to the investment bank portals cements the relationship with existing clients, and could be used to get new business. Rory McKenna, Head of R&D at DrKW explains: "Our IT strategy is based on providing personalised information to any client, anywhere in the world over any device. This can be white labelled, so a trade entry screen could be offered to a retail brokerage, thereby capturing their full order flow. Fundamentally, we are using IT for competitive advantage: our customers will use us because we are so easy to work with".
Money, Money, Money
A multi-channel offering may make sense, but what is so important about the wireless channel? Most desk-bound traders would not see any particular advantage from a wireless offering, although globalisation is changing this. The shift from country to sector trading means that traders are increasingly on duty all day, rather than when their local market is open. So even traders may see an advantage in having access to prices, analytics, research, news and transactional facilities when they are out of the office. However, at DrKW, one of the biggest successes is in corporate finance. McKenna explains: "Being able to walk onto a client site armed with equity research and contact details, and being able to access your CRM system without having to connect and log in has proved very popular". Although there are no direct financial benefits attributable to mobilising corporate finance teams, the investment required is small compared to the fees at stake and the cost of producing the research.
Raymond James, the US-based financial services firm, is also using wireless to empower its sales force. They have teamed up with 724 Solutions, a specialist wireless solutions provider for the financial services industry, to provide wireless brokerage services to their 4,300 financial advisors. The advisors will be able to access customer information and eventually trade stocks over any wireless device or network. The advisors see the big advantage of wireless as being able to securely access clients' balances and holdings, and trade stocks whenever needed, regardless of location.
Other firms do not see a business case for wireless at all. This may be because it is very difficult to separate out the benefits delivered specifically by wireless within a multi-channel offering. An interesting variation on the multi-channel theme is provided by Fidelity, which leads a very large field of brokerages offering wireless services to their retail customers. In the view of Joseph Ferra, SVP at Fidelity Online Brokerage, wireless is not just a channel. He believes it could be the dominant way in which the public conduct financial services activities in the future. The Fidelity Anywhere service offers equity trading, quotes on demand, portfolio balances, order status, news, charts and basic analytics - albeit hosted rather than running on the wireless device - for the RIM 950, Palm devices and WAP phones.
Irrespective of whether the future for wireless is single channel or multi-channel, any organisation planning to implement wireless services should have a clear vision of the value that the services will deliver. However, a business vision alone is not enough, because of a myriad of technical difficulties inherent in developing and deploying wireless services.
Mama Mia
Peter Lankford, product director for Reuters Trading Solutions Architecture, believes that financial professionals are still waiting for really useful wireless applications. "Most wireless applications to date have been simple request-response for a price quote or limit minding. We haven't yet seen richly functional wireless applications". Lankford focuses on one reason for this: the lack of support for reliably delivered streaming data to wireless devices, something that Reuters are incorporating into their Internet Finance Platform. Lisa Diamond of the Reuters Finance Innovation Group concurs with this view. "Current capabilities are still too primitive to be useful today for the professional trader. Both the tools and the bandwidth are not quite there". Diamond's team at Reuters have been working with Compaq, Sun and others on an initiative aimed at addressing these limitations. "We are focusing on four things", Diamond explains. "Designing displays suitable for small screens, providing reliable delivery of data to and from the mobile device, prioritising data over the air and adapting the service according to the user's context: the device being used, the quality of the connection, and the time of day."
Knowing Me, Knowing You
Florian Schmidt of PA Consulting expands this theme. "The real question", says Schmidt, "is how can investment banks provide their clients with filtered, relevant information presented in a suitable format when and where they need it". Schmidt believes that the key is the provision of intelligent content personalisation coupled with multi-channel delivery. Just like the Reuters vision, Schmidt sees the need for a detailed profile of the client's information needs that is dynamically updated in response to his or her interaction with the system. "It must be intelligent enough to understand the user's interests and filter, condense and disseminate the information accordingly", says Schmidt.
Based on the system's knowledge of the user, the filtering tool should be intelligent enough to automatically recognise the kind of information required. It does that not only by using specific flag-type criteria (e.g. "send me an alert when Sun Microsystems shares fall below $20") but by being able to interpret qualitatively what kind of information the user would like to receive. For instance: "I want to know about any news that could affect the price of my Sun shares adversely"; this could be a press release mentioning a slump in PC sales which does not explicitly mention Sun itself.
Apart from filtering the relevant information from various content sources, it should also automatically condense that information in a way that is based on its knowledge of the user's interests. For instance, a lengthy news article covering a number of companies could be reduced to a paragraph that contains information only on the company and issues of interest to the user.
Finally, it is important that the level of filtering and condensation can be adapted depending on the user's preferences and chosen delivery channel. A fund manager who is known by the system to be on holiday, and who has access to a PDA, may want to receive news alerts only if they are extremely important to him, and presented in a highly condensed format suitable for a portable device.
The core of any such solution would be trainable pattern-matching algorithms that are able to perform filtering and context-sensitive summarisation tasks. Software to do this is now available off-the-shelf from companies like Zapper Technologies in the US and Autonomy in the UK.
Take A Chance On Me
WAP has been heavily adopted by the brokerage and investment banking industry, primarily for share trading applications for their retail customers. This first wave has not been particularly successful, according to most industry observers, because of the "form factors" of the WAP phone, notably the small display and lack of keyboard. The suggestion that wireless computing provides a Martini experience - "anytime, anyplace, anywhere" - is generally ridiculed. Try reading a screen while driving, or writing an email while standing on a crowded train. Also, the limitations of cellular networks - slow speeds, and the need to connect to the network for each transaction - have hampered progress. According to Ian Rosarius, UK Managing Director of Macalla Software, "None of these are actually anything to do with WAP, which is actually a good network protocol." However, he concedes that current WAP devices will not support significant trading events. "They are fine for the retail market and a small subset of professional trading events, such as corrective trades or trades driven by notifications. However, large fund manager trading activity is very information intensive." Fund managers need news, research, price history, analytics and reliable real-time data, and possibly a view of market depth. None of this can realistically be displayed on a WAP phone, however ingenious the display design.
For professional users of wireless services, the form factors may be solved to a degree by the latest PDAs such as the Compaq iPAQ with its large colour screen, coupled with well designed displays. For data entry, foldaway keyboards for PDAs are already available, although these effectively turn a PDA into a table top computer, thereby reducing their mobility. On the display side, advances in handset technology could enable handset displays to be as rich as PC displays. Nokia are reputed to be working on a wireless handset that projects its display onto a surface, thereby eliminating the need to grow the handset in order to grow the display. Californian company Inviso have invented a "microdisplay" that effectively offers the equivalent of a 19-inch desktop display on a one inch screen using advanced optical technology. The screen is best viewed 3 cm from the eye - perfect for the packed train journey to work. The service is currently available on Inviso's bespoke eCase handheld device, which is currently a standalone PDA but could in principle be extended to offer wireless capabilities.
Can You Hear The Drums, Fernando?
Voice should not be forgotten. As the wireless industry searches for a "killer app", some observers have commented that it's been there all along: talking and listening. This is acknowledged by Ian Rosarius when describing the Indications of Interest (IoI) application that Macalla developed for DrKW. "Indications of Interest are really a simple information service for DrKW's clients. Users of the service are offered a navigable list of IoIs being offered. This is perfect for WAP, because it is designed around hierarchical menus. However, when you find the trade you want, you don't use WAP. You phone your broker."
WAP is both a network protocol, describing how data is carried over the cellular network, and a language for formatting data: the Wireless Markup Language (WML). WML is a variant of the extensible markup language (XML) which allows "self-describing" data. Rather than applications being expected to know the format of messages they receive, they can understand the format because it is tagged. For example, instead of receiving a message in the format "MKS 250 251 251", the XML equivalent could be:
<stockquote>
<bid_price>250<bid_price>
<offer_price>251<offer_price>
<last_trade>251<last_trade>
The push for inclusion of VoiceXML in the WAP standard recognises that giving instructions and receiving information by voice is the most convenient form of interaction for users who are truly on the move. As the name suggests, VoiceXML is another XML variant; it is used to describe structured voice interactions that you might expect from a sophisticated Computer-Telephony Integration (CTI) system. An example of how this could work is for broker research. Personalised research headlines could be delivered to wireless devices (as text or sound) with full details available in audio. This would effectively offer personalised radio programming to finance professionals on the move. This type of service is already offered by XY Network in the UK over BT's Genie wireless portal. In their words, "Forget Text - dial and listen on the move, it's what the phone was invented for". XY provide bite-sized audio reviews, entertainment plus news from Reuters. WAP is used to navigate between stories, with the sound currently delivered over a standard voice connection. Unfortunately, with existing 2G networks, the user must effectively dial up for each story. This will change with the arrival of GPRS and 3G networks.
Waterloo
As Oscar Wilde probably never said, the great thing about standards is that there are so many to choose from. This is particularly true of wireless, where there are effectively two parallel battles underway. The first is for the network protocol, which defines how data is delivered to the wireless device. In theory the network should provide reliable delivery, security, authentication and "non-repudiation" - the certainty that both parties engaged in a transaction are who they say they are. In practice, networks provide these services only to varying degrees. The problem is that most wireless networks have been designed for voice traffic. Wireless devices are connected to a cellular voice network using WAP or the Japanese "i-Mode" standard, or to dedicated data networks such as the CDPD (Cellular Digital Packet Data) network. Satellite delivery is also available, and over short distances (basically within buildings) two new standards have emerged: Bluetooth and the excitingly named 802.11.
The second battle is for control of the wireless platform: the blend of operating system, browser, markup language (the format of the data delivered to the browser) and application environment that allows applications to run on a wireless device. The devices themselves have evolved from two different lineages. Mobile phones have always been wireless but not, until the arrival of WAP and i-Mode, had the capability to handle data. PDAs and laptops have always had the ability to run data applications but have not, with the exception of a few niche devices, been connected to a wireless data network. Because of these histories, the following platforms are now jostling for dominance:
- The simplest platform is the micro-browser. Micro-browsers are used by both WAP and i-Mode. Most micro-browsers present a reduced version of HTML; i-Mode uses cHTML (c for compact) and WAP uses WML. Micro-browsers are perfectly suited to devices with little processing capability such as mobile phones. Their limitation is that they do not provide a really engaging experience for the user. A micro-browser does not provide any application processing capabilities, so all processing must be carried out on the server with results carried over the network. With connection speeds of only 9.6 kbps this can be frustratingly slow.
- An alternative platform is offered by the PDA vendors. Palm sell their own Palm operating system, which is now licenced to other PDA manufacturers such as Handspring. Psion use the EPOC operating system. Both platforms were originally designed for the personal information management features of PDAs, and have been extended to enable wireless computing.
- Java, developed by Sun Microsystems, provides application processing on the wireless device using very little memory. The Java 2 Micro Edition (J2ME) is a variant of Java targeted at wireless devices. It can run with as little as 128k available memory.
- Finally, Microsoft offer their Windows/CE operating system for wireless devices. This is a cut-down version of the full Windows operating system, which to a large extent duplicates the desktop environment on the wireless device. Wireless devices running Windows/CE are branded "Pocket PCs".
The Winner Takes It All
It is possible that platforms and network protocols will converge, making the job of the service provider simpler in that they will not need to design multiple variants of applications. The most likely convergence is around a browser-based platform, using an HTML variant, with Java providing the opportunity for local processing. Alternatively, convergence may not occur at all. The RIM Blackberry, very popular in the US but relatively unknown elsewhere, initially offered email and personal information management (contact lists, calendar etc) but not wireless data applications and not voice. It is possible that we will continue to see niche products like the Blackberry continue to appear, further complicating the task of wireless service providers.
Because of the lack of standardisation of platforms and network protocols, service providers must choose which standards, and therefore which customers, they will support. Lisa Diamond of Reuters comments: "We don't think it is realistic to provide robust mobile services over all network protocols to all devices, so we are likely to focus on a subset of standards. Windows/CE on a powerful processor like the Compaq iPAQ, running a full Java implementation could be the way to go. Getting an industrial strength version of Java, running on Windows/CE, will be very important. We don't think the slimmed down version of Java - J2ME running on KVM - will be powerful enough".
Ian Rosarius takes a different view. "DrKW asked us to build a technology platform to deliver services to employees and clients on current devices, but evolving to new devices as they arise. We try to be device neutral. Our approach is to use the same XML content and render it for different channels. We currently support WAP browsers on Palms, WAP on phones, Internet Explorer on Pocket PCs, plus cHTML for some Sony phones. All services are delivered from a common URL; our Mobility software detects the device type and renders appropriately. We have even taken the same applications and deployed them on i-Mode."
Macalla avoid the naïve approach of "transcoding" - providing a different style-sheet per device type. "The classic example of why this doesn't work", says Rosarius, "is how a portfolio of stocks would be treated. On a PC, a portfolio display would be very rich, showing earnings figures, trade histories, valuations, risk metrics, including lots of graphics. It's just not appropriate to put this onto a small screen. You'd get a big scrolling beast". Instead, Macalla use Java Server Pages to "rebehave" the content; the actual content that you get varies depending on the device you are using. In the portfolio example, tables and graphs would be summarised and presented in a format appropriate to the receiving device.
Voulez Vous
In summary, developers of wireless services should address five key questions before embarking on any development:
- What service guarantees are needed, and how can these be delivered?
- How will network resources be prioritised in the event of insufficient resources being available for all?
- How will the service adapt to cater for changes in the user's state, such as when the user is asleep?
- How will the service be tuned to fit the specific form factors of the receiving devices?
- What devices and network protocols should be supported, now and in the future?
By answering these questions, and defining a clear business vision for the services being offered, we believe that 2001 could be the year of wireless computing. Then again, we said that last year as well.
Andrew Murphy is a managing consultant at PA Consulting Group in London. Comments and enquiries about this article are welcome. Contact andrew.murphy@pa-consulting.com.
With thanks to super troupers Caroline Reynolds, Thomas Perotti, Alison Kent and Brian Rohan.
