To identify what the right mix of components and operational improvements needed to increase sales and maximize revenue for the engineering firm, the SVM team analyzed the detailed commercial transaction data of three major countries:
Australia, Singapore and Korea for an engineering company.
The dataset was cleaned up and compared against the key fields and data. From the transaction dataset, the components which had a strong association were identified using the Apriori algorithm, by identifying the support, confidence and lift value of the Antecedent and Consequent
components for various components.
In summary, the likelihood of components X and Y being bought together was identified, rather than the likelihood of just buying one component.
After a detailed study the SVM team recommended the list of components that needs to be
packaged together, which would enhance the sale and maximize the revenue.
Projections for different scenarios were simulated and the positive impact on the sale of product mix was identified. For locations where the component details were not provided, the components were identified using NLP text analytics.
Further analysis was done in the area of reducing the quotation preparation time. It was
identified that a delayed quotation was impacting the success of the quotation acceptance
from clients. A detailed NLP study of the previous Job order, such as Data cleansing, one-hot
encoding, stop word removal, unigram and bi gram word creation etc. was performed.
The SVM team created a model which predicted the quotation value when any description of
job and components used was provided. The accuracy of the model was very close to the
actual quotation, the model helped the client to dispatch quotation on time and were able to
increase the quotation acceptance.