Technical Support

From DALSolution experiences: analytical tasks are determined by many factors

Analysis is a challenge by itself

Nobody denies the usefulness of analysis. Nevertheless, managers of some large companies sometimes think that analytical services are less efficient than a clever and resourceful expert who knows “what it takes.” In countries where money is here today and gone tomorrow, and one’s success is determined exclusively by connections, knowledge of the key figures and ownership of keys to the “powder room”, analysis is not imperative. But the honeymoon is soon over, and managers eventually embrace the possibility of regular data analysis. New circumstances make analysis a necessity where it seemed to be an extra burden only yesterday. At the same time, the demand for tools with transparent mathematics and understandable analytical methodology is increasing in scientific research.

DALSolution helped elucidate customer preferences in a supermarket

Checkout records from a supermarket were examined. DALSolution helped to find which of the largest suppliers A, B, or C, were most popular among buyers. Company C turned out to be the indisputable leader, followed by company A, and finally – company B. Products of companies A and C were most commonly bought in combination (40%), followed by a combination of products from B and C (33%), and from suppliers A and B in only 11% of the cases.
Based on these results, products were arranged in the following order along the typical customer’s route through the supermarket’s retail area: first products of supplier C, then of A, and, immediately at the checkout – of company B. This layout minimized the time clients spent searching for particular products.

Insurance business: clients need time to develop

“Insurance is, primarily, a culture of lifetime planning”, says insurance company A’s customer services manager. “We work in a market where a permanent demand for insurance services is not yet established. We have used DALSolution to analyze preferences of potential clients. The results showed that the following rule is confirmed in the majority of cases: those who already have insurance coverage, are most interested in insurance. Having a relatively high rating among insurance companies, we decided to inform potential clients not only about our company, but also about insurance services in general, while at the same time conducting an effective image-building ad campaign. A more detailed analysis of these investments is in order and we plan to conduct this with DALSolution as well.

DALSolution finds customers who most need bank cards

In a Russian branch of a large commercial bank, the database of corporate clients was analyzed to identify areas of business with the largest expected increase in the number of bank card users. DALSolution found the following areas: retail trade, wholesale trade, and construction services. These three categories together represent 65% of the expected increase in the number of corporate bank card users. This information was used in the planning of bank service packages.

DALSolution helps win elections

During the 1996 presidential campaign in Russia, histograms with “For” and “Against” columns were a weekly source of excitement for TV spectators. Three months prior to the elections, Lebed’s electorate was half as big as that of Yeltsin’s: approximately 6.5% as opposed to 13%. However, a candidate’s chances are also measured by the size of the so-called “broad electorate”, i.e. those voters who will vote for the candidate in the first, second, or third order of preference. Analysis with DALSolution showed that Lebed did not concede to Yeltsin in this respect. Broad electorates were of approximately equal size, around 20%, with only a slight advantage for Lebed. This information was used in charting electoral strategies.
Another example. During an election campaign for city mayor, the analysts from a candidate’s election committee, using DALSolution, identified the following groups of voters as campaign targets: 1) aged 20 to 35, with a college education, 2) aged 20 to 49, with a high school education, and 3) aged 80 and over, with education below the high school level. The group age boundaries were computed with the help of DALSolution, based on the condition that over 50% of each group should be composed of potentially active voters who are as yet undecided about their candidate. The candidate whose staff used DALSolution, did win the election.

DALSolution helps in solving personnel problems

A metallurgical plant with several thousand workers became a DALSolution user because of personnel problems. The plant’s departments of personnel and sociological services are constantly conducting public surveys. Without DALSolution, the staff of these departments would have spent much more time on routine data processing and analysis. Their motivation was simple: promptness in personnel decision-making saves money.

Monitoring the rules of financial market

Short-term forecasting rules of stock price dynamics have a high turnover rate. Reliable rules of the form “if this occurred today and that occurred yesterday (the day before yesterday, etc), then tomorrow you should expect a particular development” certainly cannot apply to all market sectors or during different time periods. However, estimates assisted by DALSolution show that on average at least 25% of market developments can be consistently predicted.
Here is an example of a rule that held for the New York Stock Exchange (NYSE) between January 1994 and June 1998:

“If today the Dow Jones index has not risen and the relative change in the price of General Electric stock has been between –0.02972 and –0.007299, while yesterday the relative change of GE stock was between –0.01278 and –0.009631, then tomorrow General Electric stock price will not go up”.

This rule was confirmed in 11 cases out of 11, i.e. with probability 1. DALSolution was used to compute the bounds on the stock price dynamics. So far, the promptness of such computation is acceptable for data analysts working in their offices. However, this approach is applicable for monitoring rules and forecasting in real time during actual trading.

A wrong medical diagnosis threatens a pilot’s career. DALSolution helps to avoid errors

Ultrasonic diagnostic devices by Toshiba and Aloha simplified periodic medical testing of civil aviation pilots and made it more efficient. But here is the problem: during visual analysis of kidneys, even an experienced doctor may interpret a particular symptom one way or another. One diagnosis allows a pilot to continue flying, the other requires him to stop flying. An error means either the unjustified end of a pilot’s career, or the unacceptable risk of passengers’ lives. The ultrasonic survey results were analyzed with DALSolution. The software computed the bounds on kidney characteristics which allowed to tell one diagnosis from another. In all observed cases where the computed rule was applicable (there were 38 cases) it was confirmed 100% of the time without error, and it explained over 80% of all cases with this diagnosis.

With DALSolution, drug treatment is assigned more accurately

When determining treatment for medical patients, it is important to consider how similar patients have responded to similar treatment. Without modern analytical tools a considerable portion of this experience is lost. A Moscow hospital analyzed medical histories of women suffering from negative symptoms of menopause. The study was initiated by a pharmaceutical company producing therapeutic hormonal drugs that are applied in such cases. The rules found with the help of DALSolution clarified the relationship between drug effectiveness and the diseases that woman had before treatment began. As a result, the company acquired drug-specific knowledge, while the hospital could better treat patients by considering their medical history.