We work with clients in various industries. Our target industries are characterized by a large volume of data (either transactional, third-party or operational data). A list of our target industries is given below. Click on your industry to look at detailed solutions.
Our solutions span analytics (Sales & Distribution, and Operations), BI and reporting, and analytical data management. A complete list of our solutions is given below. Click on any solution to get more detail.
|
|
|
|
|
|
|
|
|
|
|
|
Our analytical solutions help consumer goods marketers to identify market opportunities to grow their business and provide inputs to allocate resources judiciously to get higher return. Click on any solution below to get more details.
|
|
|
|
We help retailers improve their store/channel performance by engaging customers through better merchandizing of the right products at the right time at the right place. Click on any solution below to get more details.
|
|
|
|
We help banks offer relevant financial services to customers with appropriate credit management and retention strategies. Click on any solution below to get more details.
|
|
|
|
We help insurance companies monitor fraudulent behavior, align quotes to risk, and offer the right product to the customers through the right channels. Click on the solutions below to get more details.
|
|
|
|
We help telecom & technology companies acquire, manage and retain customers by identifying the right offer at the right price. Click on the solutions to get more details.
|
|
|
|
We help media and entertainment companies distribute their products at the right price at the right location . Click on the solutions to get more details.
|
|
|
|
We help relationship-oriented businesses acquire, manage and retain customers by selling the right products at the right price. Click on the solutions to get more details.
|
|
|
|
Customer segmentation
Knowledge Foundry specializes in customer segmentation research for consumer and business markets. Most of our work is with psychographic (attitudes), needs and behavior based segmentation. We also work with geographic and demographic segmentation.
Segmentation should help answer some of the questions given below:
A good segmentation must do the following:
Our preferred statistical approach for customer segmentation is latent class analysis and traditional cluster analysis. We also work with CHAID analysis, CART analysis and principal component analysis for some segmentation.
Most segmentation work requires a range of sophisticated statistical techniques-factor analysis, discriminant analysis and clustering procedures. The key is to identify homogeneous groups of customers/products that have similar characteristics within a group vs. customers/products in other groups.
Our goal is to develop segments that are understandable, relevant, technically sound and actionable.
Knowledge Foundry specializes in customer segmentation research for consumer and business markets. Most of our work is with psychographic (attitudes), needs and behavior based segmentation. We also work with geographic and demographic segmentation.
Segmentation should help answer some of the questions given below:
![]() |
Can I identify subgroups of customers who may have different patterns of attitudes towards my |
| product/service? | |
![]() |
Can I understand pattern of differences among consumers on what is important to them in our |
| product/(s)? | |
![]() |
Can I identify groups of customers with similar needs? |
![]() |
Are segments consistent from region to region? |
A good segmentation must do the following:
![]() |
Improve targeting and cross sell activity. |
![]() |
Help develop deeper understanding of customers. |
![]() |
Improve marketing and communication through generation of richer profiles of customers |
![]() |
Identify white spaces, in terms of new growth opportunities for brands in the portfolio |
Our preferred statistical approach for customer segmentation is latent class analysis and traditional cluster analysis. We also work with CHAID analysis, CART analysis and principal component analysis for some segmentation.
Most segmentation work requires a range of sophisticated statistical techniques-factor analysis, discriminant analysis and clustering procedures. The key is to identify homogeneous groups of customers/products that have similar characteristics within a group vs. customers/products in other groups.
Our goal is to develop segments that are understandable, relevant, technically sound and actionable.
We work with clients trying to analyze the funnel conversion rates and RoI of conversion at each funnel stage (from online lead from a paid source to landing page visitor to buyer). We specialize in business problems where companies want to mesh demographic data with online behavior to create a better targeting strategy (e.g., online universities, reward programs) and in business cases where the conversion from landing page visitor to buyer involves offline steps (e.g., application, service rep call).
Our analysis allows companies to prioritize paid lead sources and value their RoI. It also allows companies to prioritize online.
Knowledge Foundry believes direct campaign analytics as a part of customer life cycle management process for an organization. The purpose of direct campaign is to facilitate management of all phases of customer life cycle by identifying 'Who' element for critical business questions like
![]() |
Customer acquisition: Who are likely to become a good customer? |
![]() |
Value over time: Who are likely to consider our other current offerings, next offering? |
![]() |
Retention and attrition: Who is more likely to churn? |
Our direct campaign analytics helps organization to either develop a new campaign or improve on existing campaigns. We have expertise in various types of campaigns like acquisition, cross-sell, up-sell etc. through various mediums like email, catalogs, coupons etc. We cover complete analytical component of campaign management process i.e. campaign design, campaign monitoring and campaign performance assessment.
We use various databases like transaction, customer, external/lifestyle databases, census of population and market research data to develop relevant models. Our technical expertise can deal with various issues like low response rate, small sample, spurious data, multi-stage modeling etc.
Finally, we believe direct campaign analytics have immediate and direct implication on marketing ROI.
Attrition analytics is important because of two reasons:
![]() |
The amount of money required to acquire a new customer is typically many times higher than that required to keep a customer happy and engaged. |
![]() |
Customers that are likely to attrite usually indicate their likely attrition in their usage/ purchase trends and their attitudes towards the company. However their indications are usually lost in the sea of customer data. |
Analytics helps companies identify valuable customers that are likely to churn and identify likely causes of attrition. Effective retention program targeting is a process that has certain key steps:
- Define attrition as it involves many nuances
- Pick key variables for the analysis
- Build a predictive model for attrition and rank customers on attrition risk
- Segment customers by value and attrition risk and develop retention programs for target value-risk segments
Loyalty programs help companies develop strong relationships with their customers. Good programs enable the company to increase the share of the purchase wallet in categories served by them. Typically loyal customers drive 50-80% of sales for companies.
We help clients maximize the potential of their loyalty programs through analytics. Our offerings help clients to:
![]() |
Understand loyal customers |
By employing multi-dimensional segmentation schema, we provide the client with knowledge about who the loyal customer is, what they buy and how they shop. Segmentation can be based on life-stage, lifestyles, purchase behaviors (recency, frequency, value, price sensitivity, product basket), responsiveness to marketing and attitudes. We may use these dimensions individually or combine them up to provide a more holistic view of customers along with loyalty indices/scores for each customer.
![]() |
Develop a more 'customer-centric' approach |
The purchase behavior of loyal customers can be used to make the client more customer-centric. Typical business problems include:
- Which coupons should we send to each customer to maximize incremental sales/ margin (within budget constraints)?
- Whom should we send our spring/summer catalog (e.g., family, men's, women's, youth, and kids)?
- What was the incremental sales and margin impact of last month's coupons/ catalog by segment? What worked and what didn't?
- Which customers should we invite for in-store events (e.g., wine tasting, book signing, exclusive pre-promotional reviews)?
- What bundling discount offers will work best with the loyal customer segments?
- What products are popular with specific segments of the loyalty base? Can we launch store brand products to cater to needs?
Knowledge Foundry specializes in customer segmentation research for consumer and business markets. Most of our work is with psychographic (attitudes), needs and behavior based segmentation. We also work with geographic and demographic segmentation.
Segmentation should help answer some of the questions given below:
![]() |
Can I identify subgroups of customers who may have different patterns of attitudes towards my |
| product/service? | |
![]() |
Can I understand pattern of differences among consumers on what is important to them in our |
| product/(s)? | |
![]() |
Can I identify groups of customers with similar needs? |
![]() |
Are segments consistent from region to region? |
A good segmentation must do the following:
![]() |
Improve targeting and cross sell activity. |
![]() |
Help develop deeper understanding of customers. |
![]() |
Improve marketing and communication through generation of richer profiles of customers |
![]() |
Identify white spaces, in terms of new growth opportunities for brands in the portfolio |
Our preferred statistical approach for customer segmentation is latent class analysis and traditional cluster analysis. We also work with CHAID analysis, CART analysis and principal component analysis for some segmentation.
Most segmentation work requires a range of sophisticated statistical techniques-factor analysis, discriminant analysis and clustering procedures. The key is to identify homogeneous groups of customers/products that have similar characteristics within a group vs. customers/products in other groups.
Our goal is to develop segments that are understandable, relevant, technically sound and actionable.
Market structure analysis aids clients in product/brand positioning by generating insights about the current positioning of a product and its potential with respect to competition.
The analysis helps to:
![]() |
Illustrate how competitive products/brands are perceived with respect to their strengths, |
| weaknesses, and similarities. | |
![]() |
Map existing product/brand image. |
![]() |
Assess new product positioning or the repositioning of existing brands. |
![]() |
Measure the effectiveness of communication and advertising messages. |
Some of the business questions answered by market structure analysis are:
![]() |
What are the strengths and weaknesses of my product/brand relative to competition? |
![]() |
Which products compete directly with my product? |
![]() |
Which products compete with each other and which of them are substitutes? |
![]() |
How should I reposition my product? |
![]() |
Which aspects of my brand's image are most important to consumers? |
![]() |
How can I reposition my brand to impact market share? |
![]() |
What products are most liked? |
![]() |
What new consumers can I target for my product? |
![]() |
What opportunities exist for extending the product line or introducing totally new products? |
Knowledge Foundry uses perceptual maps to explore the data and to generate hypothesis that help answer client questions. We primarily work with three types of perceptual maps:
1. Preference maps illustrate consumer preferences for a set of products
2. MDS maps illustrate product competitiveness
3. Correspondence maps to analyse and two way and multi way table
Key Driver Analysis is used for measuring the impact of relevant attributes and underlying factors on the overall view of a product or service. The analysis aims to identify critical brand, product or service attributes that have the most influence on customers' behaviors or attitudes (likelihood to purchase, overall rating, likelihood to recommend, overall satisfaction, brand equity etc). It can be used for several types of research i.e. brand preference, customer loyalty and satisfaction etc.
The purpose of key driver analysis is:
![]() |
To determine which attributes are key drivers of the particular event of interest i.e. behavior |
![]() |
To measure the relative impact of each attribute on the event of interest. |
![]() |
To identify the relative strengths of different attributes. |
At Knowledge Foundry, key driver analysis aims to answer the following questions:
![]() |
What drives overall rating for a product/service? |
![]() |
Which attributes are most important to brand equity? |
![]() |
What is the link between media spends and market share? |
![]() |
What kind of volume can I expect if I increase my distribution by 15 points? |
![]() |
Which brand equity statement is the most important in explaining purchase intent? |
![]() |
What has the strongest influence in customers returning to my store? |
![]() |
What drives overall customer satisfaction? |
![]() |
How are key attributes different across segments? |
Knowledge Foundry employs a variety of basic and sophisticated techniques to provide answers to client questions in this area.
Correlation, factor analysis and regression based approaches are primary baseline approaches used .
Advanced level analysis depending on the problem uses partial least squares, structural equation modeling, latent class regression and hierarchical bayes.
Brand marketers face several issues in improving the effectiveness of their marketing activities generally measured as sales value, volume and volume share. Few of those issues are,
![]() |
How much should we spend on marketing each brand/ line/ sub-brand? |
![]() |
Where (regions, retailers) and when (season) should we spend? |
![]() |
How should we implement each marketing input (tactics, flighting vs. pulsing, type of promo)? |
![]() |
How should we split the spend across marketing inputs (TV, print, promos, distribution)? |
Knowledge Foundry Marketing Mix Decision analytical solution helps marketers to optimize spend allocation, improving execution and testing alternative scenarios. The objective of the Marketing Mix Decision is to
![]() |
Optimize marketing spend allocation: How much to spend? Which marketing inputs to spend on? | |
| Eg: | ||
![]() |
Which brands/sub-brands provide the highest return per dollar spent on marketing? How should we | |
| distribute the budget across brands/ sub-brands? | ||
![]() |
How does a dollar spent on various marketing inputs (e.g., TV, radio, print, OOH, discount | |
| promotions, distribution reach) impact the top-line? How should the marketing spend be | ||
| redistributed across these marketing inputs? | ||
![]() |
Improve marketing execution: How to spend? E.g., | |
![]() |
What type of TV copies have the highest sales impact? | |
![]() |
How many GRPs should we target per week? How should we spread GRPs across day-parts? | |
![]() |
Which trade promotions cannibalize the least and create low post-promo dips? | |
![]() |
What is the scalability potential of Outdoors, Media Max Insertions, Print, advertorials, | |
| Interactive and Email campaigns? | ||
![]() |
Simulate top-line impact : 'What-if' and scenario analysis. E.g., | |
![]() |
What will be the impact of moving 50% spend from 30 seconders to 15 seconders? | |
![]() |
What is the Sales impact of increasing share of shelf/ Distribution ACV by 1%? | |
Our solution provides unique from the point of view that it is open box instead of black box - the statistical techniques and models are shared with the client. Our experiences with fortune 100 companies help to provide very objective and unbiased results and service.
At Knowledge Foundry we know that pricing decisions are at the top of the list for both manufacturers and retailers. Our pricing analytics provide answers to certain key pricing questions related to:
1. Price Elasticity
a. What are the price elasticity levels for my brand/SKU for different retail stores/accounts, regional markets and at a national level?
b. To which competitors is my product most vulnerable to losing share, if my competitor were to reduce prices?
2. Price threshold and gaps vis-à-vis competitors
a. Where do the threshold price points for my brand lie?
b. How much of a price premium can I charge?
Knowledge Foundry uses the non-linear log-log sales model to estimate own and cross price elasticities. We also look at promotion features in the model i.e. features, display, TPR along with store and week.
While the classical regression approach is used to generate price elastcities, we strongly recommend using bayesian shrinkage methods and latent class analysis to improve estimates of own and cross price elasticities while incorporating variances across stores, retailers, segments and markets.
For evaluating price thresholds and gaps vs. competition, we utilize sales velocity analysis. To understand price thresholds, sales velocity at varying non-promoted price points across stores is analyzed. For price gaps, sales velocity at varying gap price points between a target brand and a competitor are plotted and analyzed. These two analysis are simple and powerful ways to find best price points for a brand/SKU.
Knowledge Foundry works with retailers and manufacturers of electronics goods to recommend pricing of products based on their features. The methodology involves analyzing store sales data of various product variants with different feature levels. We use a panel regression technique to analyze pricing elasticity for each feature at a store panel level.
Retailers and Consumer Goods marketers need to address several questions to improve their promotional effectiveness. Few of such questions are
![]() |
Which product should we promote, in which stores, and when? |
![]() |
What promotion type and depth should we use? |
![]() |
How should we advertise/communicate each promotion? |
![]() |
Which promotional vehicle should we use? |
Our promotion analytics methodology helps to determine the net impact of a promotion by type and date/week in a holistic way, thus enabling comprehensive promotional planning for a product portfolio for various segments of store. For example, To estimate net effect of promotion, our methodology addresses the following questions:
![]() |
What products are likely to be affected or cannibalized by a promotion of a product? |
![]() |
What are the implications of promotion on customer basket? |
![]() |
How much is the contribution of trend and seasonality in the volume sales? |
![]() |
What is effect of delayed purchase due to anticipation of promotion? |
![]() |
How much is the effect of stock piling behavior? |
![]() |
How much is the gain from other competing brand? |
![]() |
What promotion results in cannibalization and how much is the size of cannibalization? |
![]() |
Depending in product nature, how our end-of-season sales and stock clearance different from rest of |
| the promotions. |
Our methodology promote partnership with client as each step in the methodology has distinctive output which is shared with the client, in other words 'there is no black boxing'.
Knowledge Foundry promotion analytics capabilities help marketers and merchandisers to fine balance art and science of promotion through pre-promotion optimal scenario simulation and post-promotion evaluation. We also provide customized report, dashboards or score card preparation based on information dissemination needs.
Market Basket Analysis is process of analyzing transaction level data at its most granular level-the market basket of each customer that shops at a store. The analyses of this data helps give an understanding of how the shopping happened by telling us about the quantity of items purchased in the basket and what other items were purchased in conjunction with these items.
A well done market basket analysis must offer client solutions to:
![]() |
Target offers more precisely |
![]() |
Attract more traffic to the store |
![]() |
Match inventory to customer needs by customizing layouts, assortments and price. |
![]() |
Develop more profitable advertising and promotion program |
![]() |
Increase the size and value of the basket |
Our Market Basket Analysis helps you answer the following questions:
![]() |
When and where will discounts make a difference? How do I plan promotions in the store more |
![]() |
effectively?. |
![]() |
Which items must I always stock? |
![]() |
Which products and offers will get customers into stores? |
![]() |
Which stores do better than others on demand for certain products? |
![]() |
WWhat more can my loyal customers buy? |
Our approach to market basket analysis is twofold and attempts to look at the basic analysis along with it's deeper insights-
![]() |
Use of association rules to analyse basket data and see it's size, contents and affinities |
![]() |
Deeper analysis of patterns i.e. at store, time of day, multiple visits, campaign or promotion level to |
![]() |
uncover more valuable insights |
We also recommend validating rules across different segments to provide a richer insight to the analyses.
Relative allocation of shelf space for various SKUs is a critical driver of category sales. Our methodology involves segmenting stores by sales and sell-through across products. We then analyze sales data across store segments with different shelf space allocation to recommend changes in allocation.
E.g., for a specialty mattress retailer with multiple product sub-categories, we recommended the relative allocation of space at various store segments.
For a Category manager, it is a constant challenge to identify, and make available to the consumer, a set of items in his category that would maximize his returns/increase the brand's market share/improve the brand value, or any such objective. A lot of factors contribute to this multi-criteria decision making scenario --- shelf space constraint, budget constraint, product penetration, consumer perception of product, etc., to name a few. It is very likely that in the whole mix, there are a few items that are not "efficiently managed", in the sense that they seem to be giving sup-par returns on the investment made.
In such cases, where one needs to identify and address relative "underperformance", Data Envelopment Analysis (DEA) can be used. In the DEA methodology, developed by Charnes, Cooper and Rhodes (1978), efficiency is defined as a weighted sum of outputs to a weighted sum of inputs, where the weights structure is calculated solving a linear program for each decision making unit (DMU).
In the context of category management, one could treat each category as a DMU, and define appropriate inputs and outputs. For instance, the inputs or resources used could be shelf space, number of SKU's, price discount information, product feature information, etc., and the desired outputs could be the category's total sales, profit margin, basket share (% of tickets that had at least 1 item from that category), perceived variety from supply perspective (importance of feature/attribute) and demand perspective (market share covered by the category's SKU's)
Such an analysis can be done at {Store, Category} level at a monthly level to generate valuable insights on how to optimally utilize the resources to attain the desired objective.
Coming soon...
It is always challenging for financial institutions to identify "optimal" locations for setting up their branches. Typical break-even period of a new branch can vary between 12-24 months, and hence defining "Expected branch profitability" becomes difficult. Since one also needs to take into consideration, the population stability in terms of financial and demographic characteristics used to assess if a particular ZIP/county/district is "potentially profitable" or not, the scenario becomes even more challenging.
A practical approach:
We first begin with assessing the existing branch network, i.e., conduct a detailed performance analysis of every branch. We obtain a list of all indicators that the bank bases its branch financial performance evaluation on. Some examples of indicators are - the number of customers, value of deposits, personnel expense, operating expense, deposit interest expense, value of delinquencies, interest income from loans, credit cards and other sources, return on assets, etc. These indicators can be classified into one of the two categories - resources or services. Essentially, one needs to measure a branch by its ability to convert the resources at its disposal into services valued by the bank management.
There are several techniques to evaluate the branch performance. One widely used methodology is Data Envelopment Analysis (DEA). DEA is a linear programming technique for measuring the relative efficiency of decision making units (DMUs) where each DMU has a multitude of desired outputs or needed inputs. DEA helps in rank ordering the branches by evaluating their ability to minimize resource utilization while providing a given amount of services, or in terms of their ability to maximize service provision from a given amount of resources.
Further, each region (ZIP/county/district) needs to be evaluated on various dimensions such as:
![]() |
financial need, capacity, risk appetite, |
![]() |
historical response rates to various offers, |
![]() |
geo-demographic profile, and |
![]() |
other economic factors like presence of competitors, retail outlets, etc. |
Typically, the new branch location is identified based on a lot of business criteria. A few of them are cited below:
![]() |
Is the resource level at the existing branch sufficient to cater to the trade area's current demands? |
|   | Does the trade area offer business potential (from results of evaluation described earlier) to warrant an additional branch for that region? |
![]() |
Does the proposed location meet the "minimum distance from existing branch" criterion? |
![]() |
How does re-assignment of regions affect existing branch dynamics? Are there sufficient prospects in |
| the newly carved trade area to sustain business? |
Knowledge Foundry helps clients select channel partners/distributors/ agents. Industry verticals such as telecoms, technology, insurance, wealth management have extensive third-party databases capturing detailed information about channel partners/ agents.
We select channel partners based on sophisticated business rules on the data fields available e.g., Channel partners with >50% sales from Small business, with Revenue >$10MM, with Salesmen >10, based in a particular region, who are not company X channel partners. Very often we have criteria that balance the selection of channel partners across geographies, verticals or customer size to ensure uniform market coverage.
We also segment client's existing channel partners based on existing performance metrics (available with the client) and based on third-party database information.
We also help clients create Partner Dashboards that score channel partners on parameters such as:
- Target vs. achievement
- Relative Growth
- Frequency
- Linearity (across weeks, months, quarters)
- Percentage of Lead Conversion
- Reps trained on company's products
- Exclusivity
- Promo program participation
- Tenure
We offer fraud analytics solutions to clients in insurance, healthcare, and retailers. Some of our solutions include:
Inconsistency/ misrepresentation analysis (Insurance): For P&C Insurers, we analyze the variables in insurance applications, to highlight applications which may contain misrepresentation (by the applicant or by a broker/ agent). We do this using a fuzzy algorithm that examine the consistency of information in fields such as age, income, profession, education, home ownership status, coverage requested, driving history claimed, etc. We also have techniques to detect multiple applications from a computer for direct insurers and examine changes between these applications to detect 'fishing-for-a-better-rate' behavior.
Claims pattern analysis (Insurance, Retail): For P&C Insurers and Health Insurance companies, we mine data patterns in the claims data to find abnormalities and inconsistencies in claims data. Abnormalities include high incidence or severity of claims in specific regions, by type of claim, or by a specific service provider or customer. Inconsistencies include high labor rates, medical charges or part charges for standard procedures. We also analyze claims severity and loss ratios by the characteristics of the asset insured to detect possible misrepresentation.
Check fraud and gift-card fraud analysis (Retail): For retailers, we analyze patterns in transactions involving checking fraud to identify which regions, banks, and products bought have a higher probability of fraud. We also develop and maintain blacklists of customers that have committed checking fraud or gift-card fraud.
We offer a range of solutions in credit and collection analytics, primarily to clients in telecoms, casinos and credit cards. Some of our solutions include:
Credit scoring: We help credit card issuers, casinos and telecom companies decide the credit limit for their new and existing customers. For new customers, we help companies create Application Scorecards ('A' scores) based on statistical analysis of customer socio-economic profiles and credit report details. For existing customers, we help companies create Behavioral Scorecards ('B' scores) based on statistical analysis of customer profile, credit report details, and transaction history in the past 6-12 months. These scores are used to determine the credit limit assigned to the customers.
Delinquency models: We help clients predict which customers are likely to be delinquent in the next 3-6 months. This is done using predictive modeling based on customer socio-economic profiles, credit report details, and transaction history (outstanding balances, payment history, delinquency status).
Collection treatment optimization: For clients with delinquency exposure to customers, we develop tools to predict what collection treatment (E-mail, gentle letter, phone call, collection agency) works best to recover monies from delinquent customers in each time horizon (15 days due, 30 days due,.). We use predictive modeling based on historical transaction behavior (outstanding, past due days, recovery behavior), customer profile and credit details.
Our actuarial analytics solutions include:
![]() |
Comparing mortality/ disease rates to actuarial tables: We calculate mortality rates and disease |
| by demographic groups of age, gender, smoking / alcohol consumption, genetic history, and prior | |
| medical condition) and compare these rates with actuarial tables. |
![]() |
Analysis of the factors driving differences in mortality/ disease rates: We use statistical techniques |
| (such as decision trees and regression) to explain differences in mortality or disease rates using age, | |
| gender, , smoking / alcohol consumption, genetic history, and prior medical condition as independent | |
| variables. This helps the insurance company understand the factors correlating with higher mortality/ | |
| disease rates. |
We offer a range of analytics solutions in supply chain and logistics. Some of our solutions include:
![]() |
Forecasting: We have experience at forecasting demand rolled up at various levels of the bill-of- |
| materials(raw material, sub-assembly, ., final product) by sourcing location using appropriate | |
| forecasting horizons e.g., For a jeans manufacturer we forecasted demand by SKU (up to 9 months | |
| out) and rolled that back into 'cut-level' forecasts (4 months out) and 'fabric-level' forecasts (9 months | |
| out) | |
![]() |
Inventory management analytics: We have worked with manufacturers in setting ERP parameters |
| (safety stock, reorder level, batch/order quantity)to optimize total costs based onstatistical principles. e.g., for an industrial products company, we set the safety stock, EOQ, and reorder levels | |
| at their warehouses and plants in the USA and Mexico | |
![]() |
Logistics analytics: We have experience in using logistics data (locations, travel times, routes, cost |
| for LTL/FTL, capacities, product shipment characteristics) and demand forecasts for route scheduling, | |
| asset scheduling and warehousing/ distribution optimization, using statistical techniques. E.g., we | |
| helped a manufacturer optimize their own-fleet routes by reducing empty backhaul miles |
We offer a range of analytics solutions to aid procurement. Some of our solutions include:
Spend Visibility Analysis:We analyze key direct and indirect spend categories of the client to determine the composition of the client procurement budget by sub-category, by location, and by vendor. Very often, spend visibility analysis is preceded by a detailed data cleaning effort which involves classifying invoices/spends which are categorized as 'unclassified' spend.
Vendor selection analytics: We help companies in selecting vendors for procurement by helping them create a harmonized vendor RFP based on past spends, and comparing vendors on cost, timeliness, quality and other metrics. E.g., for a transport-intensive manufacturer, we helped reduce courier costs by creating a harmonized express shipping schedule, and helping them choose courier vendors based on their rates for route-service level- weight combinations.
Vendor compliance analysis: We help companies ensure that their vendors are invoicing them in compliance with their contract. This involves checking pricing used on invoices, checking bulk discounts and trade discounts and ensuring accuracy of labor hours charged (for service vendors). E.g., we helped a retailer and a CPG manufacturer validate charges for trade discounts and other types of discounts
Structured reports and dashboards
Many companies have information residing in multiple data sources (e.g., internal data sources such as transaction data, loyalty program data, survey data; and external data such as panel/ scanner data or company profile data from third parties or retail store data from retailers). It is critical for decision makers to get an integrated view to assess the health of the business. Knowledge Foundry helps clients create structured reports and dashboards in the following areas:
![]() |
Brand health dashboard:To track the market share (by region, retailer), awareness, media | |
| exposure (by medium), share of voice, % ACV - by week/ month/ quarter/YTD and YoY | ||
| growth | ||
![]() |
Distribution health dashboard: To track the sales performance of the company | |
| distributors/ agents/ brokers by product/ business unit. For relationship-oriented businesses, | ||
| distribution dashboards could include churn trends, lifetime value of customers acquired, and | ||
| overall book of business contributed - by week/ month/ quarter/YTD and YoY growth | ||
![]() |
Inventory dashboard: To track sell-through (and inventory days) and inventory levels by | |
| product at key locations - by week/ month/ quarter/YTD and YoY growth |
Simulation/ Scenario and Stress test tools
Decision makers are regularly faced with questions which require analysis of the impact of changes in macroeconomic factors or competitor behavior on their business. Moreover it is critical for them to decide the optimal response of the company in such situations. Knowledge Foundry helps its clients in creating simulation tools or scenario ("what if") analyses to prepare for changes in assumptions. Some examples of such tools include:
![]() |
Tool to simulate sales with different media spending scenarios in a media mix model | |
![]() |
Simulation of weekly sales of a product category in a promotion effectiveness exercise using | |
| different promotion scenarios | ||
![]() |
Scenario analysis to test the potential revenue impact of a pricing change (using different | |
| elasticity scenarios) |
Ad-hoc SQL queries
The IT teams of many companies do not have the time and resources to respond to ad-hoc business questions from decision makers, which require queries on the company's databases. Our analysts can write customized SQL queries to obtain the desired information. Sample questions include:
![]() |
Which loyalty program members from the mid-Atlantic region did not buy anything from us | |
| last quarter? | ||
![]() |
Which loyalty program members live in a 5-mile radius of Zip Code 10178? | |
![]() |
Which loyalty program members spent more than $10,000 in the past 12 months and bought | |
| a digital photo frame from us? | ||
![]() |
What were the total sales of Apple i-Pods in each of our stores in the mid-West this quarter? |
Business intelligence tools
Knowledge Foundry analysts can structure standard reports in the client's BI system (Business Objects, Cognos, Hyperion) to generate critical business reports for decision makers on a periodic basis.
We cover all aspects of analytical data management, which involve the following activities:
![]() |
Data cleansing: Ensuring that the data is treated for outliers, mis-formated information, junk data, | |
| and missing values | ||
| may have mis-matches (e.g., Phonetic matching, rule-based matching). This allows the company to | ||
| have a single view of a customer across databases. | ||
![]() |
Data preparation: We have deep expertise in creating and transforming raw data variables into | |
| analytical variables e.g., time shifted variables, derived variables, transforms | ||
![]() |
Maintenance: We ensure that the above steps are carried out in a regular schedule to ensure the | |
| quality and consistency of the analytical data. |
In changing economic times, the accuracy of forecasting is critical for the success of any business. Knowledge Foundry has significant experience in forecasting using both time-series and cross-sectional time-series data.
In our past work we have experience forecasting over multiple time horizons to support different business objectives (e.g., in the apparel industry - fabric forecast, cut forecast, assembly forecast, shipping forecast). We also have experience in forecasting at different geographic roll-up levels (store, region, store-chain, national).
At Knowledge Foundry we understand that our clients need to manage risk while maximizing revenue on a regular basis. Our Risk Analytics Solutions provide clients with the ability to measure manage and mitigate risk along the customer lifecycle.
Acquisition Models: We build application scoring models that allow clients to take on board only those customers who as applicants have a low predicted probability of future default. This helps in bring in customers who are profitable and filters out those that are not.
Behavioral Scoring Models: Our behavior models help the client assess the profitability or risk of a customer by looking at his entire credit transaction history along his life cycle. The models employ robust statistical procedures to predict future default by looking at payment patterns of customers. Our models help in-
- Analyzing transaction history and predicting payment patterns
- Preventing default and reducing delinquency and collection losses
- Segmenting customers based on their risk levels and identifying opportunities for growth
- Setting appropriate credit limits for customers
- Optimize cross sell and upsell opportunities
Some examples of our behavior models are-Collection scorecard models and delinquency models.
Apparel