Global communications technology company saves over $5M using AgenaRisk
A leading technology company, known around the world for innovation and leadership in wireless and broadband communications has achieved annual savings of over $5m by using AgenaRisk to calculate the reliability and increase the quality of new electronic products.
The ubiquity of technology in business and everyday life, increases the pressure to get products to market as quickly and efficiently as possible, to steal competitive edge and demonstrate cutting-edge technology. At the same time, it is vital that each product goes through all the necessary quality control measures to avoid disrupting customers' business continuity. This is especially critical within the Network and Infrastructure Division of this global communications company which delivers applications to support mobile phone lines and government and military networks.
Previously the Company used a range of methods to predict the hardware reliability of their products, including accelerated life testing methods, which simulate the ageing process, and those found in US DoD Military Standard 217 (MIL-HDBK-217). However, they found that accelerated life testing was very costly and extremely time consuming and that using the methods recommended by MIL-HDBK-217, were inaccurate - they simply came up with the wrong answer.
The Company felt that they had to break the mould of reliability testing. They had no confidence in existing methods because component level analysis didn't offer the predictability they needed. Instead they found that reliability of components tended to be much more random than expected, or significantly worse than reported by industry sources. In addition they faced the common industry problem that many returned products, when analysed, were not found to be faulty. From this they concluded it was not possible to make a reliable calculation of risk using traditional methods.
Starting from scratch, the Head of Quality spearheaded a new approach, segregating failures by archetypes with different types of field failures. Using standard statistical measures, the group started to find correlations and built a model to capture this. Recognising they were on the right path, a decision was taken to find a technology partner to help them put in place the right solution to predict risk in a more intelligent, automated process.
After conducting extensive research they decided to use Bayesian Networks as the basis for its solution and worked with Agena to implement a new approach to quality control and reliability prediction. Agena's risk analysis and decision support software tool - AgenaRisk - uses all the information that traditional business intelligence tools use, but then combines this with innovative Bayesian networks and advanced predictive analytics.
The Company recognised that Agena understood that there were better ways to make more accurate assessments of risk. Using the power of Bayesian networks and AgenaRisk, they realised that the Quality Control team could focus on the critical process variables, enabling them to more accurately predict delivered product quality and do so at lower cost. Whereas previous methods relied on detailed quantitative data, often of unknown provenance, AgenaRisk could mix quantitative data with qualitative data in the form of engineering judgements about vendor quality, process maturity and system complexity. This made it possible to completely change the Company's approach to predicting risk.
Agena consultants worked closely with the Quality Control team and other company consultants in the UK and US to design, test and deploy the new model within AgenaRisk.
- Flexibility of Bayesian Tool allows it to be adapted according to changing business needs
One of the first advantages which the Company has realised is the flexibility of the AgenaRisk model. Agena worked with the Quality Control team to ensure the model was built to be generic, based on inputting critical data points which are the most 'predictive'. The result is a very flexible tool which can reliably predict risk, even when adapting to changes in the business and to subsequent new requirements.
All business intelligence tools require good quality business data to ensure the end result is effective. The Company identified the information which is most important, looking at the area of potential impact as well as the potential risks.
One significant benefit of using Bayesian modelling is that it is very flexible. The Quality Control team don't need all the facts to get the end result. For example, if an area of work is outsourced and you lose access to certain measures, you can still run the model and it will work. As long as you know the critical variables, you can still predict the impact accurately.
- Return on Investment Potential Impacts Future Strategy
The Company has developed an analysis of the financial impact of using AgenaRisk. By using Agena's risk assessment tool to prevent problems before a product gets launched into the field, initial figures demonstrate a strong return of investment.
There are three main ways to reduce costs by launching products of high quality. First, if you launch a product of higher quality it means a reduction in return rates within warranty. Second, you can target test more precisely so there are either less people involved in the testing process or testing is simply quicker - either method saves money. Third, the number of design staff with the expertise to resolve customer problems can be reduced. This frees up their time to focus on their main role - developing next generation products. All of these have a big impact on potential costs.
CONSEQUENTLY THE POTENTIAL SAVINGS ARE PROJECTED TO BE OVER $5 MILLION IN 2007 ALONE. THIS INCLUDES A 30 PER CENT IMPROVEMENT IN WARRANTY EXPENDITURE.
- Business Impact of Successfully Predicting Risk
Other business units within the Company are just beginning to realise the business opportunities which can be achieved by using AgenaRisk. With confidence in the results running high, the newest version of AgenaRisk was trialled for the launch of an important new product. Although it is too early to measure its success, the Quality Control team were able to present back on the initial findings, highlighting the areas where the new product needed to be revised. On the strength of these findings, the Quality Control team have had corporate approval to use AgenaRisk on the launch of all new products.
Using AgenaRisk has led to a complete change in mindset within the Company. Figures from previous control measures provided measures of potential reliability. But this never changed launch plans; it simply warned the Company of fault issues. Now, when the Quality Control team present figures from AgenaRisk the model dynamically tells the team what to do to meet the quality control checks and people within the Company are acting upon these results.
Since the Company has given the green light to use AgenaRisk more extensively within the business, there are already a number of plans for the future. These include incorporating AgenaRisk in the initial architecture of product designs to ensure higher quality at a far earlier stage.
As the Head of Quality concludes: "Even the cynics are beginning to see the potential of AgenaRisk. Recognising that you can accurately predict risk and prevent problems is changing people's minds about how to view the business. AgenaRisk is now driving our risk mitigation strategy. This is a very powerful tool."
Ed Tranham, Commercial Director
T: + 44 20 7404 9722
32-33 Hatton Garden London