About us

Company background
Clients
Case studies
Client testimonials
How to find us
Biographies
Vacancies

Case Study: Web Load Test of Hertfordshire Council schools admissions

Around 178,000 children are educated in Hertfordshire's network of 550 schools and in each year around 30,000 children are involved in moving on to the next stage of their education - from nursery to primary and then on to secondary school.

Hertfordshire schools are among the best in the country and many are oversubscribed. In allocating places, the council follows a series of published criteria such as a child's special needs. More than 90% of children receive a place at one of their three preferred schools, one of the best rates in the country. Despite this, the schools admissions process - from when parents choose a school to discovering whether their child has got a place - is inevitably a sensitive time for both parents and children.

Hertfordshire County Council actively encourages the public to access its services online, and was the first in the country to run a full schools admissions service via its website. Set up in 2003, the online system allowed parents to nominate their three preferred schools in order of preference, and following the allocation process by the council, discover which school their child had been allocated a place with. In the first year everything ran well during the initial nomination phase and take-up online was excellent, until the secondary results came out. At this point - a matter of minutes, the council's web systems were inundated with users the moment the results were published, and the service for parents and children at a very sensitive time was not of an acceptable standard.

To address this, in 2004 the council put in place a new caching system - a big step-up in capacity from the previous year - but still the website became overwhelmed as parents all seemed to want to access the site at midnight when the results came out.

BBC local TV were following a number of families going through the admissions process, some of whom were applying online, and therefore the problems the website was facing were broadcast across the region. Adverse publicity would undermine the council's objective of encouraging the public to use the internet to access a wide range of its services and not merely schools admissions.

In preparation for 2005, the code that delivers the results to parents after login was re-written. But would that be sufficient or should the council go to the considerable expense of adding extra hosting capacity?

For 2005, "Hertfordshire just had to make it work" states the council's E-government Manager "as any bad experience can destroy public confidence not only over admissions but across the board for other council services. We wanted to test 'live' to ensure that we were testing the actual technology. We didn't want to test to destruction as we needed the website to stay in service throughout the tests." To provide the answer, the council brought in web effectiveness testing specialists SciVisum to run independent Load Tests to determine just how many users per minute could be processed online after the upgrades.

SciVisum's approach involves scripted up User Journeys - multi-page routes through the site to reproduce exactly how parents use the site on results day. Some might start at the special eAdmissions URL sent to them, or they might start at the Council's homepage and browse the site to search for the results. Hundreds of dummy parent accounts were set up to allow many logins in parallel.

Testing proved that the changes made by Hertfordshire's IT Department for 2005 were indeed performing well and that the rewritten results-delivery coding was sufficiently robust. Moreover, it demonstrated that the costly option of an extra hosting facility was unnecessary, saving the council several thousand pounds.

An additional and unexpected benefit was that the testing pinpointed possible bottlenecks as yet not considered by the council. Homepage user journeys demanded too much server resource. The available bandwidth was insufficient to handle the same volume of users as the results server.

With less than a week to go, the council team were able to address these two areas; including the construction of a new lightweight homepage to be used on the big day - one which would be both quick to deliver and have very clear links straight to the results section.This would reduce the number of pages for parents navigating to the results. The proof of the pudding came on the evening of Tuesday March 1st 2005 at 5pm when the secondary admissions were announced. Several of the IT team were on hand in case anything should go wrong. But all went well.

"it was so important for the eAdmissions service to run smoothly and we are grateful for the help provided by SciVisum to resolve the problems.The site was completely rebuilt after earlier problems. Load Testing not only assured us that we had overcome previous problems, but also highlighted new areas we would not have spotted. After the bottlenecks of the two previous years it was essential that public confidence would not be undermined."

"We are very happy to be a reference for SciVisum.They did an excellent job for us.We saved a considerable amount over a rather expensive extra hosting facility that had been offered to us and we now know that we have a good platform for future admissions - whether secondary, junior or infant."

Graph of Web site traffic in the 90 minutes after results announced

Additional Information

Will it Work? Six key steps to Stress / Load Testing - newsletter August 2004
November 2004 - Leading fashion brand selects SciVisum to stress test e-commerce site to cope with unpredictable shopping sprees.
Hertfordshire Council School admissions web pages
eAdmissions.org, The eAdmissions National Project (UK)