Having worked on both sides of the fence, I am always interested in how software vendors connect with their customers, and how customers view software vendors. In theory, there should be an intimate connection: vendors need customers, not just for revenue but for references to help gain more customers, and so you would expect them to keep good track of how well customer implementations of their software were going. At the very least you would expect software vendors to regularly survey customers on how happy they are. The reality may surprise you.

I recently conducted research in which I asked software vendors to supply customer references to which a short satisfaction survey would be sent. You might think that this would not give very much insight, since the vendors themselves nominate the customers to be used as references, and how hard can it be for a vendor to find a chosen set of happy customers? It is rather like putting personal referees on your resumé: you should be able to find at least two people to say something polite about you.

In fact, the exercise proved surprisingly revealing. For a start, given that each vendor had agreed to contact the customers beforehand to check they were happy to participate in the survey, you might expect that the references provided would at least be current. In fact, a considerable percentage of the reference emails bounced. Vendors seem not to have been smart in selecting only happy customers either, as there were a surprising number of moderate scores, and some distinctly less than complimentary comments from buyers.

This little exercise highlights what I believe is a major systemic problem in the enterprise software industry. Vendors are driven by their investors to hit their sales targets, and the compensation structures used for sales forces drive them to complete the next big sale. At this point, often after a lengthy competitive evaluation process, the vendor sends an invoice, the software is shipped, and the attention of the vendor moves to the next sale. Sure, there will be some ongoing contact from the professional services organisation of the vendor, who will probably deliver some training and perhaps help to implement the first project using the software. However, the attention of the vendor is likely to fall off after that, coming back into focus around the time that the customer is due to pay their annual maintenance fee.

The larger the company, the more disconnected it is likely to become from its customers. A market research firm called Nucleus had some fun in the late 1990s by looking up the public customer references of some large software vendors on their websites, then contacting those companies and asking them to take a customer satisfaction survey. One survey of ERP customers found that not one vendor could manage a score of even 50 per cent in answer to the question: "Would you recommend this software?", which is a pretty basic level of customer satisfaction for what were supposed to be public reference customers.

Certainly some vendors do a much better job of keeping up with their customers than others, and in particular companies who work on an annual fee basis (rather than a perpetual licence model plus annual maintenance) have a particularly keen incentive to ensure that their customers are reasonably tolerant of their software. Privately-held SAS Institute, which uses a renewable annual fee rather than perpetual licences, is noted for its very high renewal rates (well over 90 per cent), and I recall someone who worked there saying that one key to this was the standard contract required that a company give 12 months' notice of cancellation. This gave the vendor a year to try and identify where there were problems and go out of their way to fix situations that had gone awry. In many cases problems were turned around, and the customer eventually renewed.

Many software salesmen are in their nature hunters rather than farmers, and enjoy the thrill of the chase. Studies show that it is dramatically cheaper to sell a second product to an existing customer than it is to acquire an entirely new customer. A sale to a new customer costs four times as much as a sale to an existing customer (some studies claim six to seven times, or even more). Surely it would make sense for the software vendors to worry more about how well their existing customer implementations were going?

You do not want a stagnant user base, but the effort typically devoted to developing ongoing relationships versus acquiring new ones is not optimal. Software companies have much to gain from closer relationships with existing customers. The thrill of flirting with a new love may be exhilarating, but putting in the hard work of making an existing relationship work may pay greater dividends in the long term.