A01 chaf6542 06 se fm indd

Chapter 12  Digital business service implementation and optimisation

are posted to our blog 

www.smartinsights.com/blog

 and then we use tagging of topics in 

different categories to surface the advice in relevant sections of the site. Each post is syndi-

cated to Google’s Feedburner service using RSS feeds which can then be automatically posted 

to Twitter. We use Hootsuite to post to other social networks like Facebook, Google+ and 

LinkedIn so that a more conversational tone can be used to recommend content or ask ques-

tions than is possible simply through updating a social network automatically. Hootsuite can 

also be set to automatically track posts and the impact of sharing using social media, using 

campaign tracking to add parameters to the URL of the post. The end of the cycle is where 

email is used to share content weekly with an editorial written to recommend the most useful 

posts or major updates for our readers to be aware of.

Feedback from users and analytics will reveal different types of errors and ideas across 

a website. A change request system is needed to prioritise and manage these, so that the 

 highest-  priority issues are corrected first. We can apply the fault taxonomy of Jorgensen 

(1995) to an e-commerce site to decide on the timing of the action required according to the 

type of problem. We can see that the approach is quite different from that for a traditional 

information system or packaged software that is distributed to thousands of customers. For 

example, with a mild problem such as a spelling mistake within software, it would be too 

costly to update and redistribute the software. With the e-commerce site, a spelling mistake, 

although trivial, can be updated immediately by correcting it on the web page or in the data-

base or content management system where it is stored. Indeed, minor problems need to be 

corrected because they reduce the credibility of the site.

For more major errors, it is essential to fix the problems as soon as possible since revenue 

will be lost, both from customers who are unable to complete their current purchases and 

from users who will be unprepared to use the site in future because of their bad experience. 

Data from transactional e-commerce sites show that very few have continuous availability. 

If the site revenue for a 24-hours, 7-days-a-week site is £10 million per week then if avail-

ability falls to 95% this is the equivalent of losing £500,000, before the loss of future revenues 

from disgruntled customers is taken into account. A modular or  component-  based approach 

to e-commerce systems should enable the location of the problem module or cartridge to 

be identified rapidly and the problem in the module to be fixed, or possibly to revert to the 

previous version.

As well as fixing the problems shown in Table 12.3, companies will also wish to update the 

functionality of the e-commerce system in response to customer demands, sales promotions 

or competitor innovations. Again, a  component-  based approach can enable  self-  contained, 

discrete, new modules or cartridges to be plugged into the system which are designed to pro-

vide new functionality with only minimal changes to existing modules. For each such update, 

a  small-  scale prototyping process involving analysis, design and testing will need to occur.