£200m unaccountable in school's ICT procurement

Becta cannot account for over £200m of taxpayer's money spent on software for schools according to data obtained under the Freedom of Information Act.

The scheme known as e-Learning Credits (eLCs) was set up to make funds available to schools to buy software from an approved list of titles and suppliers.

By April 2006 schools had received £300 million pounds of funds to buy 'educational software' intended to enhance the curriculum. Operating systems and Office suites were specifically excluded from this scope.

Becta reported on eLCs in the CAB Report to The Secretary of State, Report Number 3 December 2005 and it was the source data for this report which forms the basis of our commentary.

The data is incomplete as a quote from the FOI return from Becta makes clear:

"Detailed analysis of eLC returns suggests that they continue to understate the actual level of expenditure in the market. In particular, we have found that eLC returns appear to underestimate the real level of sales of those who have provided returns by as much as 70% and that there is a significant number of suppliers, some relatively large, who do not submit eLC returns".

The Data

Over 14,500 software titles are listed by Becta:

• 27% is classified as "Drill and Practice" and are largely aimed at younger students (KS1);
• 11% of the titles relate to assessment software;
• 7% is for "exploration" software;
• 26% is for information resources;
• the remaining 29% is not classified according to use or age group.

The eLC returns are for the period January to September 2005.

Commentary

Given the above limitations we have the following observations to make on the data supplied:

1. The lack of eLC returns means that, in effect, over two thirds of the expenditure with accredited suppliers is not accounted for.
2. Nearly one third of listed, approved titles are not classified by Becta making judgments as to their suitability hard to make.
3. 70% of sales returned through accredited suppliers is to giant software vendors and to a single ICT hardware supplier to schools. The range of these companies' education portfolio combined with a single eLC return make it impossible to audit which software titles are sold to schools.
4. In most cases the transparency of supplier status is difficult to ascertain. For example 'Education Bradford' is a wholly owned subsidiary of SERCO, and '3E's Enterprises (Trading) Ltd' whose sales make up one fifth of the total are a wholly owned subsidiary of the City Technology Trust - a DfES accredited supplier for the BSF bids.
5. Over 45% of eLC sales are to the dominant ICT hardware supplier who is also the QCA’s contractor for the National Curriculum on-line assessment project. This company exclusively supplies software for only one operating system.
6. A significant percentage of software is for administration purposes (contrary to the original remit). Examples include library management software; skills assessment and (ironically) many thousands of pounds were spent on software to administer proprietary licences.
7. Where active choices by schools can be audited, the great majority of software purchased divides into: design and technology related software (KS4 /Tertiary); music and media related software (KS3-4); and literacy and numeracy software (KS1 nad SEN).

Conclusions

The ineffectiveness of the checks and balances put in place to monitor government spending on software in schools has resulted in a huge waste of public money and a distortion of the ICT market in schools.

There is the suspicion in the low eLC returns that some companies are not adhering to the Becta guidelines. This assertion supported by anecdotal evidence from schools some of which have bought hardware through eLC funds.

Major incumbent suppliers have been the chief beneficiaries of the funds and the market has been unfairly distorted. Smaller companies and free software suppliers have been adversely affected and innovation has been stifled.

The strategy of software-title accreditation in tandem with supplier accreditation is inherently flawed. No data is available to evaluate the impact of eLC software in most areas of the curriculum.

In the absence of evidence to the contrary we regard the eLC project as the silent example of another failed, expensive Government ICT project adding to an already depressingly long list.

Executive Summary

The UK is spending over half a billion pounds on ICT in schools. The DfES figures indicate that £500 million year is being spent directly¹ and our figures show £100 million is spent indirectly on electricity². In addition, projected figures based on government targets for computer and interactive white-board provision will see these figures more then double in the next few years.

Our report was prompted when BECTA (the DfES's ICT quango) made the very serious statement in May 2006 that the level of funding in schools was unsustainable⁵. We carried out an analysis of ICT expenditure for a typical secondary school and calculated the break down as follows⁴:

• Support: 65%
• Electricity: 15%
• Software and Licences: 10%
• Hardware: 10% (5 year renewal cycle)

These figures were surprising in that the level of support looks disproportionately high and the ratio of provision (hardware and software) to overheads (support and electricity) is very low at 1:4. In other words, for every new computer purchased at £300, £1200 would be added to the total cost.

We concluded from our analysis of the breakdown of costs that the way that ICT has been allowed to develop in our schools has resulted in over-complex systems being supplied which when scaled up generate unsustainable levels of support burden.

In addition, it became apparent that vendors can no longer support these systems profitably at the rates schools are able to pay. In effect vendors' profits are being underwritten by the taxpayer because they have adopted an inappropriate technical model for ICT promoted by DfES and BECTA.

The report below shows how schools differ from a standard business model and provide a set of unique challenges which need to be tackled new ways.

ICT challenges in educational settings

ICT in schools is now a peculiar hybrid of the standard business model and a pedagogic model. It is not suited well to either task.

All of the usual ICT administration tasks found in medium and large companies are found in schools but with a few onerous twists:

• Users come and go in their hundreds and change their privilege status as they age.
• There is a duty of care with regard to unsuitable mail and Internet content.
• There is a need to prevent user-initiated abuse of the system at a level that would cause mass sackings in a business environment.
• Software diversity is almost unimaginably great to reflect teaching needs.
• Most software, including mission-critical databases, are accessed almost simultaneously due to the way a school timetable and term is organised.
• The administration overhead would tax even well-resourced IT departments in large businesses let alone a staff of three in a normal secondary school.

Given these unique conditions it is essential from the outset to deploy systems which do not also carry inherently high levels of low-level but time-consuming technical support. Unfortunately this exactly the opposite has happened.

Third-Party Support Problems

Schools are simply not set up to pay the rates for third-party support common in the business sector to service conventional networked systems.

This becomes clear when it is considered that supply teachers are paid at rates of £150 a day (and feel they have done well), typical "one man band" local technical support (much deprecated by BECTA) charges £250 per day on site and INSET consultants rarely exceed a £300 daily rate.

At these prices no professional support organisation can profitably service school ICT needs, and schools will not pay for example £800 per day typical in industry. As a result, support on-site is now largely provided in-house. How this is implemented is described briefly below.

Standard model

A secondary school will have two or three dedicated technicians, one of whom at least is a qualified systems administrator. Often the school will have a "back-stop" support contract with the equipment vendor.

Deluxe model

The truly wealthy schools (mostly Independent or those state schools picked out for special status) having ICT support departments of ten people are not uncommon. It is no surprise that at this level such schools are happy with their systems, they work well and have many exciting features. One particularly wealthy school in South London supports it facilities with a staff of 20.

Costs

But salaries add up. The "standard model" set up costs over £100,000 per year in wages. The "deluxe model" will cost between £250,000 and £500,000 on ICT-related staff. The wealthier schools illustrate well that the support overhead required to maintain this peculiar quasi-business computing network is potentially enormous and the conclusion must be that the current computing model used in school is not appropriate.

Recommendations

School ICT needs a complete re-think. New facilities, more equipment and features for school ICT will just make things worse. In BECTA's own words: "a step change is needed to make schools ICT sustainable"⁵.

Money is being wasted on low-level support ICT by in house staff who spend their time on mundane tasks such as rebooting frozen computers, re-imaging corrupt hard drives, retrieving files, clearing printer queues and dealing with virus related tasks.

Our recommendations are as follows:

There should be a clear separation between administration and pedagogic facilities. Very surprisingly this is not the case in most schools where the one 'network' serves both. The administration facilities should be specified, financed, audited and supported in line with any other business model.

The sheer scale of pedagogic facilities however means that these must move their technology to ultra-low maintenance systems in order to free the schools from the technological burdens of ICT.

This means that workstations must be thin-client terminals, services where possible delivered on-line and ultimately high maintenance computers such as servers removed altogether, or at the very least serviced remotely.

Such models are available now, and in the Open Source community they are technically mature. What is needed is the political will to do things differently or indeed the situation really will be unsustainable.

References

1. The Publishers Association

2. http://www.dfes.gov.uk/pns/DisplayPN.cgi?pn_id=2006_0160

3. http://www.siriusit.co.uk/index.php?page=electricity-consumption

4. The basis for calculation is the 'average secondary school' referred to in our previous article.

In this school we have 3 technical staff, one employed as qualified administrator.

Approximate salary bill is £100,000 pa for all three (excluding on-costs).

The electricity bill to run two hundred computers as defined in the same article is approximately £20,000 pa.

Software costs and licencing are very variable within establishments but £15,000 would be a fair average. Similarly, based on quotes supplied by major vendors, workstation renewal is approximately £300 per unit.

Thus for 200 computers replaced on a 5 year cycle with a zero value write down, this amounts to £15,000 pa.

The grand total is £150,000. Thus 66% is support, 13.3% is electricity, 10% software/licences and 10% is hardware.

5. Quote from key note speech by Steven Lucey, Director at BECTA, in Coventry, 2^nd May 2006.

by Dr. John Spencer, Head of Business Development (Education)

Many measures have been produced to cost information and communication technology (ICT) in UK schools. These include capital equipment replacement cycles, the cost of software licensing and upgrades, the cost of the software applications themselves and the cost of technical support.

All of the above contribute to the total-cost of ownership (TCO) of school ICT and collectively is a cause for concern to BECTA, the government’s educational quango for ICT.

BECTA’s report on TCO released in May 2006 pointed out that savings of up to 60% are possible if schools use Open Source Software in the place of proprietary software.

Our investigation took on a different dimension when we realised that those schools with large Open Source commitments also tend to deploy on non-traditional hardware configurations.

We decided to compare the power consumption of two technical models of computing in schools, one we shall call the "Conventional Model" and the other the "Green Model".

To produce our figures we first have to define the models, detail their power consumption and specify their use.

Secondly we will base our figures on the secondary education sector as they are the biggest users of computers in education.

It is worth noting at this stage that the secondary sector is more homogeneous than the primary sector and that there are approximately 5000 schools in the UK compared to a total of 59,000 registered educational establishments.

The Models

Conventional

In a 'conventional' network the workstations are regular PCs with nominally 350w psu and with either 150w 15" CRT monitors or 50w LCD 17" panels. There are typically 5 servers per school running at 700w psu each and housed in an air-conditioned (5KW) room.

These are conservative figures many schools have more powerful desktops (400w), more servers and typically air conditioning in all network rooms. CRT monitor's power consumption vary considerably. LCD panels draw between 35 and 75 w when active depending on the manufacturer.

Green

In a 'green' network workstations are thin-client terminals based on the power over ethernet model (PoE) and typically draw between 1 and 2 w per terminal. 30 thin-clients are served by one 400w terminal server.

The 5 network servers used in the Conventional Model are consolidated into one 800w server running virtualised operating systems and are liquid cooled.

We assume the same 50w LCD monitors as in the 'conventional' network. Lower power consumption means that no air conditioning systems are required.

Parameters

Numbers of schools and computers

We worked on a basis of 5000 secondary schools (including independent schools) in England, Wales, Scotland and Northern Ireland, and 200 PCs per school. ^1,2,3,6

There is no consolidated record for the entirety of this sector but the figure of 200 PCs per school is conservative.

Cost of electricity

The cost of electricity is set at 6p per Kwhr – this is the mode cost in the UK. Local Authority schools have had electricity supplied at a discounted rate, a practice which ceases in 2007 when they will pay full commercial prices.⁷

The carbon dioxide to KWhr was set at 650g of carbon dioxide per kilo watt hour of electricity produced - this is the mode taken from power generation sources in the UK. ⁸

Typically gas-fired power stations produce 440g of CO2 per KWhr, oil-fired stations produce 650g per KWhr and coal-fired stations 950g per KWhr. At present Local Authority schools have preferential rates for electricity costs allowing them to absorb high electricity consumption. This is due to end in 2007.

Usage

A schools network servers (and required air conditioning) run 24/7, 365 days a year.

We can assume that workstations are switched on for 8 hrs per day, 5 days per week for about 0.7 of a year to allow for holidays and weekends.

Calculation of costs & emissions

Using these parameters we calculated the following figures for our notional, average secondary school.

Conventional Model

If we scale these figures up to include the full population of 5000 schools in the UK we arrive at nearly £100 million spent on electricity per annum and 1 million tonnes of carbon dioxide released.

Green Model

The Green Model therefore represents a 89% saving in the cost of electricity and a 78% reduction carbon dioxide emissions when compared to the Conventional Model.

Future projections

The trend based on the conventional model is for the number of computers per school to rise.

The stated aim of many authorities is to have one computer per child. In addition the exponential growth of the interactive whiteboard in all education sectors is set to achieve one in every classroom.

Bearing in mind that an interactive board runs from a conventional PC with a 600w projector and that there are over 50,000 primary schools in the UK we can predict a ten fold increase in power consumption with concomitant carbon increases over the next five years.

The green model proposed is essentially a model to rework the conventional model with the absolute minimum of changes in the resources available and the way in which they are used. It is worth mentioning that a third model can be deployed which takes the concept of reduced power consumption further.

A school might well decide to make use of applications and storage provided over their Internet connections and for personal computing use embedded applications either on the thin-clients or on personal usb -keys. In this scenario the school 'network' dispenses with on-site servers completely and uses only the PoE terminals. Such a setup can easily be powered by renewable sources of energy such as wind turbines or solar panels thus reducing the carbon costs to near zero.

References

1. The UK Education Disk
2. The Education Publishers Council
3. Pupils in Scottish Schools
4. Independent School's Council
5. The Publishers Association
6. DfES
7. Carbon dioxide - www.dft.gov.uk and www.esru.strath.ac.uk
8. ElectricityCosts