Phil Wainewright, one of my favourite SaaS (Software-as-a-Service) bloggers, wonders whether it’s Time for a SaaS code of conduct. He mentions another recent Gmail outage, and makes the valid point that SaaS outages tend to get much more publicity than internal corporate Microsoft Exchange outages (and “Microsoft only warranties the CD the software comes on. There’s no guarantee the software itself will actually work.”). But the different delivery methods mean that SaaS customers have completely different expectations.
When (that’s a when, not an if) things go wrong, SaaS vendors need to respond quickly to their customers, Phil says. Customers should not be lulled into a false sense of security (‘it won’t fall over; if it does, only rarely’), and when the inevitable happens, they need to be kept properly informed and reassured. He proposes a five-point code of conduct. The key points are reproduced below, with my own commentary based on experiences at [my employer] BIW Technologies:
- Spell out exactly what the contract does and doesn’t deliver – The BIW contract includes a Service Level Agreement that specifies the functionality of the BIW platform, gives monthly uptime commitments, and prescribes the levels of help desk support, information back-ups, utilisation monitoring and management, security provisions, and training, deployment and consultancy support.
- Give customers a clear plan for when things do go wrong – BIW maintains its production environment in a specialist managed hosting facility that is separate to the ones used to host its corporate website and to manage email communications. In the event of a major outage, therefore, the company can quickly provide alternative sources for information in the rare event that the BIW system suddenly becomes inaccessible. Disaster recovery plans are shared with customers so that they know how BIW will return the service to active status, and how communications will be maintained.
- Report live service level metrics – BIW constantly monitors its capacity statistics and captures and analyses figures weekly to predict growth trends. It hasn’t yet gone as far as Phil recommends in providing real-time data to its customers about their usage of its system, but this may just be a matter of time.
- Have a customer insurance plan in place – Phil reckons the SaaS industry needs a properly funded plan that allows for an orderly transfer of their data and processes to a new home. There isn’t one for the industry as a whole yet, but BIW’s contingency arrangements include both a software escrow agreement and a contingency scheme whereby its managed hosting provider, Attenda, will continue to provide the BIW service for at least a further two months in the unlikely event of BIW folding.
- Let customers leave whenever they like – BIW customers are not tied into long-term contracts. They tend to pay monthly or quarterly subscriptions for the delivery of specified levels of service, with rights to terminate the agreement and move their data elsewhere if they wish, subject to similar lengths of notice. Moreover, any moves would also be facilitated by BIW’s provision of customers’ project data in a recognised industry standard format (XML). Fortunately, such moves are very rare as for BIW – like many SaaS businesses – the key ‘S’ is not Software but Service: BIW knows that it needs to maintain high levels of service to retain its customers’ loyalty and limit ‘churn’.