IT

Eighty percent of success is showing up. — Woody Allen

My company, Kefta, helps its clients, usually Fortune 500 companies with e-commerce operations, improve their online conversion rates. We typically increase sales by 10–20%. This is not rocket science, more akin to Retail 101, simple things like modifying pages to stop showing offers for products we know the user has already purchased, or making offers more relevant when we know the prospect is interested in a specific product (e.g. because they come from Google after searching for that keyword).

Sometimes I wonder if what we are doing is not too sophisticated by far, when I see particularly boneheaded practices at places that really should know better. Dell is often touted as a model of logistical and operational excellence, and for being a web-centric company. My experience is that many products they carry are not listed on the web site and can only be ordered by phone. You also have to phone to get a discount.

Despite being a telecoms engineer by training, I loathe phones. Phones are great for keeping an emotional connection with friends and family, but are a staggeringly inefficient form of communication for business purposes. They do not leave an audit trail, and even when they do (my voice mail system automatically forwards them to me by email as a MIME-encoded WAV attachment), they hog disk space and are not searchable. You can scan an email in a few seconds, but are forced to listen to voice mail at whatever pace it was dictated. Well, at least with WAV attachments, I can skip back to write down a phone number without having to replay the whole message.

Coming back to Dell, I recently needed to buy a Gigabit Ethernet switch from them. I sent an email to my rep, which he promptly ignored. I tried calling, at least 4 or 5 times, but my only option was voice-mail jail. In the end, I passed the buck to a junior colleague, who tried to leave voice mail and discovered he couldn’t because it was full. With persistence, he managed to get Dell to condescend to taking our order. No customer should have to go through so many hoops just so the vendor can take their money.

I am ragging on Dell, but most IT vendors do as poorly. I can understand expensive support calls receiving lower priority and resources than sales calls — after all, the company already has your money. Not having their act together for the simple matter of order-taking simply boggles the mind. Workflow systems, automatic call distributors and other technologies designed to prevent this have been available for many years. It looks like nobody has bothered to go through the user experience, even though these bugs (and many other glaring deficiencies like session timeouts) could be caught by the most cursory of inspections.

Dell sends an automated satisfaction survey after a sale. Unlike the order-taking process, the survey follows up if you do not respond… That said, it is the usual worthless multiple-choice question format asking me to answer irrelevant questions on a scale of 1 to 10. I don’t recall if the form had a box for free-form comments, but even if it did, the survey design is not-so-subtly signaling that no human is ever going to read what you type there, and thus it is not worth the effort to fill it. The numeric answers are probably going to be collated by an automated report nobody pays any attention to anymore, because garbage-in, garbage-out.

If you are serious about customer feedback, make it open and free-form, and make sure each and every feedback is read by a human (they come quite cheap in the Midwest and the developing world). They should be acknowledged personally (not with an automated reply) and followed through until the issue is either resolved or a decision is taken not to implement the changes suggested (because they are too expensive, impractical or whatever other reason). In both cases, inform the user who bothered to give feedback — most large companies pay a fortune in market research while at the same time ignoring the free (and usually very valuable) insights submitted by their customers. Granted, you cannot always resolve every complaint by unreasonable customers, but feedback on process issues should always be taken into consideration.

Sometimes dropped orders are due to active incompetence rather than careless neglect. While implementing a campaign for one of our clients, we realized there was a bug in one of their ordering forms that would cause them to drop an order. Our software sits on top of the client’s website and monitors it precisely for exception cases like these, and we told them we could, at no extra charge for them, send the dropped order details to an email address of their choosing so the order could be re-entered manually. They declined our offer for various reasons related to internal politics and trade union issues, essentially they were refusing to bend down and pick up money lying on the floor (our estimate was they were losing tens to hundreds of thousands of dollars of customer lifetime value every month due to inaction).

You don’t have to endure a multi-million dollar ERP or CRM implementation to improve follow-through. Where there is a will, there is a way, and a little creative thinking will usually find a work-around that can get the job done until a more robust solution can be deployed. One of our clients, a major bank, was in the early stages of developing their e-commerce, and simultaneously in the throes of a Siebel implementation. Their online forms would simply send an email to a branch office for manual processing. We were implementing a satisfaction survey for them, and offered to send an email automatically to a supervisor if the customer’s order had not been processed, at least until Siebel came on-line. Poor man’s workflow, but email workflows are often quite effective, specially for remedial situations like these.

As I mentioned, sometimes I think I am in the wrong business, and should instead start a consultancy to teach some clue to large companies that have grown complacent. But then again, that is assuming somebody cares, beyond paying lip service to Customer Relationship Management. There is no point in setting up complex systems to build a lifelong relationship with repeat customers if you can’t even take their orders in the first place.

When I first heard Microsoft chose “Vista” as the official name for the much-delayed Longhorn release of Windows, I immediately thought of this (Quicktime, 657KB). Does this mean I am a bad person?

There is definite sense of malaise at Microsoft today. On the face of it, this is odd – the US government, while successful in demonstrating illegal behavior, only struck them with an ineffectual slap on the wrist and no monetary damages. Microsoft then settled all its legal liabilities with AOL, Sun, Burst et al for what amounts to pocket change for them. The nuclear winter in the tech industry means there is no investor money flowing to a challenger, with the possible exception of Google.

Yet the MSFT stock price languishes (that has a corrosive effects on employee morale as stock options no longer function as incentives). More worrisome (for them), there seems to be a sentiment the company has accreted bureaucratic bloat and lost its ability to deliver products.

Think whatever you want of Microsoft products, their modus operandi has always been to come out with a laughable 1.0 product, then a so-so 2.0, followed by an acceptable 3.0 and a 4.0 to mop up the competition if the 3.0 has not already. In many cases, Microsoft products prevailed not because of their intrinsic brilliance, but because of their evenness – no brilliant features, but then again no fatal flaws either. This takes a very disciplined software development process. The fact is, no fundamentally new version of Windows has been released since Windows 2000, and Longhorn keeps slipping and shedding features like integrated search that Apple, a much smaller company with a far less impressive R&D budget, manages to deliver today.

Bill Gates saw first-hand how the IBM anti-trust process, while never brought to its conclusion, paralyzed IBM’s decision-making with legal hand-wringing over compliance issues. Some pundits believe the Microsoft is being gradually enervated in the same way . As anyone who has worked with corporate lawyers can attest, they sometimes seem to think the only way to avoid liability is to do nothing, much as decapitation is a sovereign remedy to headaches. That said, Microsoft has consistently shown a pattern of brazen contempt for the law, it is hard to imagine that changing just because Bill Gates was ridiculed for his testimony.

Others, of a more curmudgeonly bent, attribute Microsoft’s inability to deliver products to their recent emphasis on blogging. Writing a half-decent blogging certainly takes time, more than one realizes because intense cognitive activity alters the perception of time. That said, it is hard to take this hypothesis seriously. Few Microsoft bloggers have the prodigious output of a Scoble, and arguably his blogging is a direct part of his job as evangelist, not a distraction.

The core of the problem is that Microsoft is still a two-trick pony. The bulk of its revenues and profits come from its Windows and Office monopolies (and ancillaries like development tools). Despite all the cash poured into diversification efforts like MSN or PocketPC, they have not managed to expand significantly outside these two cash cows. The problem is, both desktop operating systems and basic productivity suites are maturing markets that are fast becoming commodities. Having successfully fended off commercial competition, they are now left with competitors they cannot financially choke away.

Free (as in beer) software threatens not so much to destroy their monopoly but rather to put such intense price pressure on their margins as to effectively condemn Microsoft to stagnation. Electricity, gas and water utilities are also monopolies, but they do not enjoy the same pricing power or stock multiples, and that’s due to the commodity nature of their businesses, not just regulation.

The corporate market is saturating, and IT managers have learnt from the bubble that applying stable technology to business problems is much more profitable than constantly chasing the bleeding edge in search of an elusive payoff. Microsoft pulled a fast one with their notorious Licensing 6.0 program that left a sour taste in many an IT manager’s mouth, and any further attempts to gouge the business market will likely lead to widespread rebellion and wholesale moves to OpenOffice.

The consumer market is left, and that’s where I think Microsoft is in the biggest trouble. At one point Office was more valuable than Windows, but it has been stagnating of late. The dearth of compelling reasons to upgrade since at least 1997 or so certainly accounts for that, leading Redmond to clutch at straws with ridiculous ad campaigns.

I have two Office licenses at home, Office 2000 on my PC (well, it has broken down last week, and I simply can’t bring myself to care), and Office.X on my Mac. I never use either. It’s interesting to look at why:

• Microsoft Word is a program consumers mostly use to write administrative correspondence like letters to their banker – those trying to write the Great American novel are probably not a significant user base. Guess what? You don’t write letters to your banker anymore, you use their online banking site, and at worst write them an email. I haven’t used Word for personal use in at least three years. Not a good sign for the flagship component in the suite.
• Much the same can be said of Excel – it is used mostly for personal finance or as a simple database program. You don’t use a spreadsheet to manage your bank accounts or stock portfolio – you use your online banking or a more specialized program like Quicken or Microsoft Money (the latter is the only Microsoft program I still use regularly for my personal use). As Microsoft Money is now for all practical purposes a loss-leader for MSN, it is hardly going to be a savior for Microsoft’s margins.
• As for Outlook, three words: Hotmail, Yahoo, GMail. Enough said…
• Frontpage is still popular with small businesses but is declining. For personal publishing, it has been all but superseded by easy to use blogging software from the likes of Six Apart or Google.
• Powerpoint and Access are more specialized programs that are mostly used in a business setting. In this networked age, databases are meant to be shared and the kind of people who can deal with Access are more likely to set up a small PHP+MySQL app on a cheap hosted account.
• The more consumer or small-business oriented editions of Office include their also-ran photo editing suite. It faces crowded markets with good-enough free competitors like Picasa, or Adobe Photoshop Elements, the de facto standard for advanced amateurs, not to mention the myriad of so-so shovelware that comes bundled with digital cameras or scanners. Video editing is another segment where Microsoft is weak.
• Microsoft Publisher is similarly a program that has trouble breaking out of a crowded field.

In all these examples, there is one constant – most applications other than rich media creation are moving to the web. Not in the way originally envisioned by the advocates of the network computer, circa 1998 – not general productivity suites (à la Office, only ported to Java), but rather the underlying business processes themselves reimplemented in a way that does not require editing documents, and more importantly for the companies concerned, no manual processing of documents on the other end, with huge labor savings for those concerned.

So what is the way out? Perhaps the audience for software is fragmenting the same way mass audiences for other forms of software such as music or television have exploded with the proliferation of choice. Apart from video games, writing software is still a low-overhead industry with few barriers to entry, and the Internet has made distribution much more cost-effective for small software companies, specially when the product is delivered as a service over a web browser Microsoft can buy out small startups, but the opportunity costs of management attention often outweigh potential revenues. Quite possibly there is no way out of Microsoft’s (admittedly comfortable) predicament.

Somebody at Oracle clearly spent a lot of time creating a quad-precision floating package that exploits the 38 digits of the built-in Number type. Pointless, yet impressive.

SQL> select to_char(exp(1)) from dual;

TO_CHAR(EXP(1))
----------------------------------------
2.71828182845904523536028747135266249776

SQL> select to_char(4*atan(1)) from dual;

TO_CHAR(4*ATAN(1))
----------------------------------------
3.14159265358979323846264338327950288422

Dialogue from The Librarian, a TV movie on TNT. Sheltered “professional student” Flynn Carsen (played by Noah Wyle of “E.R.” fame) is kicked out of school and is interviewing for a job at a vaguely supernatural library:

Charlene (Jane Curtin, coldly): What makes you think you could be the Librarian?

Flynn Carsen (Noah Wyle): I know the Dewey decimal system, Library of Congress, research paper orthodoxy, web searching, I can set up a RSS feed…

Charlene (stifling a sigh): Everybody can do that. They’re librarians.