# 1 Intro: Project X

/// _Need to work on wording_

* 1.0 Project is a novel brand-to-consumer engagement, communication and data analysis platform that allow unprecedented insights into consumer behaviour and allows on-thy-fly direct engagement and feedback capability putting FMCG marketing on par with digital product development.

* 1.1 We offer a completely novel data tool to measure consumer behaviour that looks into individual consumption patterns directly and is retail channel agnostic. We acquire data points on **what?**, **when?** and **where?** the consumption takes place that allows us to answer the most important questions of **why** and **what next?**

* 1.2 In the **Analysis** section you can find a few sample usage scenarios that a lot of marketers and salespeople found useful for understanding their consumers, improving targeting, acquiring capability for virtually real-time feedback collection and agile experimentation that thus far was available only to digital products.

* 1.3 Unlike the vast majority of the data analytics tools and platforms, we offer significant interactive capability that allows you to tailor make your data acquisition exercises "on the go", push information to the consumer to gauge the response, run A/B tests.

* 1.4 The platform also has a fully functional sales capability that allows pushing offers to the consumers and directly measuring the results, while benefiting from the extremely precise targeting capability.

# 2 Data sources

///_Here we provide the schematic overview of the ecosystem, outlining how our data is collected_///

## 2.0 SKU Barcode scan

#### Scan results

* SKU Name
* Location

#### Post scan survey:
* Retailer/ Channel
* Price
* Quantity
* Discount information
* Sentiment analysis

## 2.1 Store visits / scan point redemption

* Items viewed
* Name
* Category
* Price
* Brand

* Searched terms

* Purchase history
* SKU Name
* Category
* Price
* Brand
* Quantity
* Discount preference

### Customer information
* Name
* Age
* Sex
* ID
* Phone number
* Street address
* p2p network

## 2.2 Anti-counterfeiting Label Scans

* SKU name
* Category
* Authentication scan location
* Authentication scan time
* Confirmation scan location
* Confirmation scan time

### 2.3 Promotion Coupons Interaction data

* Category/Item?/Brand
* Clickthrough
* Conversion

# 3 First Derivatives

///_Some examples on what one could do with the data through very simple manipulations (the list is not meant to be exhaustive but instead should provide some ideas of what could)_///

## 3.0 Consumer Profiling

* Shopping history
* Budget estimation
* Income Elasticity (Discount response)
* Item/Brand/Category

* Region, Sex, Age
* Loyalty (Fox vs Hedgehog)
* Brand/ SKU (by category)

## 3.1 Brand (SKU) Profiling

* Locations
* Retailers
* Price level
* Discounts
* Typical consumer
* City size, Sex, Age
* Budget/Income
* Share wallet
* Brand
* Category
* Brand within category
* Mean Elasticity
* Associated brands (brands most frequently bought together)

## 3.2 Regional/ Retailer Profiling

* Most represented brands
* Typical consumer
* City Size, Sex, Age
* Budget Income
* Share wallet (Retailer)

# 4 Analysis

/// _This part is meant to showcase what exactly can be done with the platform in terms of generating Actionable Insights_///

## 4.0 Case 1: Data Deep Dive

/// _Here I wanted to present a "sample analysis" of what you can do with the data that we provide to get to actionable results. Obviously we'll be refining this a lot as we interact with the counterparts and anyone is wellcome to suggest any other ways to showcase the capability_///

* In this section we'll be running a sample data analysis exercise to get to generate insights that you and your team can act upon Today. This is just sample of many different ways our platform and data can be used for your particular needs.

* 4.0.0 **Consumer Clustering** Identify different consumer clusters

* We'll start by running a smart clustering algorithm to identify distinct groups of consumers who purchase your products

| Consumer cluster | Brand's share in Category portfolio | Monthly spending |
| ---------------------------- | ----------------------------------- | ---------------- |
| I: Loyal Consumers | >50% | mean +- 15% |
| II: Repertoire High Spenders | <30% | mean + 50% |
| III: Emergent Consumers | <20% | mean - 20% |

* Clusters **II** and **III** --> **High Potential** : The algorithm determined that there are 3 distinct groups of consumers who are different from each other but whose members exhibit very similar behavioural patterns within the groups. For the purpose of this exercise we are going to focus on **Repertore High Spenders** - consumers who spend a lot but who juggle different brands and **Emergent Consumers** - those who seem to have only started purchasing items in our category.

* 4.0.1 **Preference Analysis**

* Now, we can key consumer **preference terms** unique to each cluster through consumer survey response. We use these **terms** to better understand our consumers in these clusters and to better tailor our marketing communication in future.

| Consumer Cluster | Preferences |
| ---------------------------- | ------------------------------------- |
| II: Repertoire High Spenders | quality, form -factor, availability |
| III: Emergent Consumers | variety, value, availability |
| |

4.0.2 **Competition Analysis through consumption Portfolio**

* Now we can look deeper into the purchasing habits of our consumer clusters namely what other brands in our category do they typically purchase.

| Consumer Cluster | Brand - % Portfolio |
| ---------------------------- | ------------------------------------------------------------------- |
| II: Repertoire High Spenders | Brand A - 40%, Brand B - 25%, **YOUR Brand - 15%**, Other - 20% |
| III: Emergent Consumers | Brand A - 25%, Brand B - 50%, **YOUR Brand - 10%**, Other - 15% 20% |

* We can see that within each cluster we face a different chief competitor.

* We are going to do a deep Dive into **brand A** and **brand B** offerings to the SKU level to identify key drivers of the different brand performance within the clusters.

| Consumer Cluster | BrandA-SKU % | Brand B - SKU % | Your Brand |
| ---------------------------- | ---------------------------------- | --------------------------------- | ---------------------------------- |
| II: Repertoire High Spenders | SKU1 - 10%, SKU2 - 80%, SKU3 - 10% | SKU1: 30%, SKU2 - 30%, SKU3 -40% | SKU1 - 25%, SKU2 - 50%, SKU3 - 25% |
| III: Emergent Consumers | SKU1 - 40%, SKU2 - 30%, SKU3 -30% | SKU1: 50%, SKU2 - 30%, SKU3 -20% | SKU1 - 10%, SKU2 - 20%, SKU3 - 70% |

* We can see that brand performance is so a large degree determined by the performance of its different SKUs within different consumer clusters.

* We can use the cluster's **preference terms** that we obtained before, to explain the different SKU dynamics within consumer clusters

* --> Breaking down SKU performance by **key Preferencesnce terms**:

| Brand | SKU | quality | form-factor | availability | value | novelty |
| ----- | --- | ------- | ----------- | ------------ | ------ | ------- |
| A | 1 | 60 | 50 | 50 | 40 | 40 |
| A | 2 | **80** | **90** | **70** | 40 | 40 |
| A | 3 | 55 | 50 | **70** | 35 | 50 |
| B | 1 | 50 | 50 | 60 | **70** | **80** |
| B | 2 | 56 | 56 | 60 | 40 | 60 |
| B | 3 | 60 | 60 | 50 | 40 | 40 |
| YourB | 1 | 60 | 40 | 50 | 40 | 30 |
| YourB | 2 | 65 | 50 | 60 | 35 | 35 |
| YourB | 3 | **70** | 65 | 60 | **80** | 50 |

* We can clearly see that different SKUs are evaluated differently according to the **key preference terms** which is very indicative of different product/market fit and messaging.

* 4.0.3 --> Key takeaways:

* It is worth focusing marketing activities on 2 promising Consumer Clusters: **Repertoire High Spenders**, **Emergent Consumers**
* The two clusters have different priorities -> different marketing approach should be adopted for optimal efficiency:
* Competitive **portfolio analysis** and **survey sentiment** analysis indicate that marketing strategy should be focused on:

* **Repertoire High Spenders** -

* Refocus marketing campaign by emphasising **quality** and **ease of consumption** for premium SKUs
* Increasing **premium SKU** availability through better distribution in **offline premium retailers**
* Improving **SKU2** form-factor to better accommodate consumption patters

* **Emergent Consumers**

* Refocus marketing campaign by emphasising **novelty** and **value** for value SKUs
* Redesign **SKU3** to make it more appealing to the target consumer gorups (**novelty**)
* Increasing **value SKUs** availability through better distribution in **online retailers** and **corner shops**

## 4.1 Case II: Sentiment Analysis

/// _After users scan the barcodes we can prompt them to answer certain quesitons vis a vis this particular SKUs.. This should be a "top of the mind" measurement that we can charge a premium for if someone decides to run a custom study_///

* **Category**/ **Brand** / **SKU** sentiment analysis based on user feedback

* After scanning SKUs from a particular category users are asked for a feedback. We collect and aggregate this feedback to evaluate key attributes and preference terms for particular SKUs, Brands and Users.

* Typical quesions:

* **"My main reason for choosing this product over its competitors is"**:

| Brand | SKU | value | packaging | form factor | quality | taste | availability |
| ----- | ---- | ----- | --------- | ----------- | ------- | ----- | ------------ |
| A | SKU1 | X | | X | | | X |
| A | SKU2 | | | | X | X | |
| B | SKU1 | | X | | | | |
| B | SKU2 | | | | X | | |

* In addition to the pre-existing terms that you can query in our database, you can also run live studies to gauge user sentiment. Custom made studies at this interface allows to capture top-of the mind sentiment that is a much more representative of the real decision making than any survey could ever achieve.

* Here is an example of using this matrix to identify competing SKUs as seen by the actual consumer:

* **"When this product is not available I typically go for:**

| | ASKU1 | ASKU2 | BSKU1 | BSKU2 |
| ----- | ----- | ----- | ----- | ----- |
| ASKU1 | - | | X | |
| ASKU2 | | - | | X |
| BSKU1 | X | | - | |
| BSKU2 | X | | | - |

* --> Competitive perception matrix allows to immediately gauge how your product is stacking up in users decision tree. E.g. **Brand A SKU1** and **Brand B SKU1** are perceived as direct substitutes whereas **BSKU2** is not in the consideration set of other products in the category implying a weaker competitive position of **BSKU2**

* The platform's interface is sufficiently flexible to accommodate a large variety of different studies.

* Examples of **Tailor made push messages**:

* same category scan --> **"would you consider [insert yours] product?"**
* same category scan --> **Click here to get a coupon for a discount**
* same category scan --> **rate [your prodcut] packaging/form-factor/availability/quality**
* same category scan --> **Did you know?? [your product] info**

## 4.2 Case 3: Surveys

///Y.B: _Here we can offer access to our consumer who want to earn extra cash and don't mind spending time filling out the questionnaires". Typically it takes brans around a months to get something like that from Nielsen. I feel there is alot of opportunity in providing this kind of data very rapidly_ ///

* Test alternatives and gather user feedback through rapid surveys [survey campaigns to earn points pushed to user]

* **Fine tuning marketing campaign**:

* Consumer targeting criteria:

| Gender | Age Group | Region Size | Region Location | Income Group | Brand Loyalty Index | Category size in Portfolio ｜
| ------ | --------- | ----------- | --------------- | ------------ | ------------------- | -------------------------- |
| Female | 25-40 | 2,3,4 Tier | South | 5000-10000 | 50 | Major |
\

* Which words do you think describe the product best?

| "pure roots" |
| ------------------- |
| "rose water" |
| "flower power" |
| " amino moist" |
| "super transparent" |

* Which shade of pink do you prefer?

| img1 |
| ----- |
| img 2 |
| img 3 |
| img 4 |

* Which animal/plant do you prefer

| "cat" |
| ----------- |
| "dog" |
| "snake" |
| "honey bee" |
| "rose" |
| "daisy" |

## 4.3 Case 4: Demand Prediction and Dynamic portfolio data

/// _With a sufficiently large core of regular users we'll be able to measure dynamic changes in the "share of wallet". Speed and rapid feedback might be key selling points for this feature_///

* Our regular and engaged user base allows us to track changes in the **share of wallet over time** as well as to gauge the effect of marketing/sales action virtually in real time.

* Below is an exapmple of a report of for average **promotion discount** for a category as it changes over time:

* Brand(SKU)/ average discount in a region over time series

| | | I | II | III | IV | V |
| ------- | -------- | --- | --- | --- | --- | --- |
| Brand A | | | | | | |
| | Region 1 | 12% | 15% | 17% | 19% | 20% |
| | Region 2 | 10% | 9% | 10% | 11% | 8% |
| Brand B | | | | | | |
| | Region 1 | 10% | 5% | 6% | 7% | 13% |
| | Region 2 | 10% | 10% | 10% | 10% | 10% |

* We can see how **brand A** is pursuing an aggressive trade-discount campaign in region A

* Brand(SKU) share of wallet in a region over time series

| | | I | II | III | IV | V |
| ------- | -------- | --- | --- | --- | --- | --- |
| Brand A | | | | | | |
| | Region 1 | 5% | 6% | 8% | 12% | 15% |
| | Region 2 | 12% | 13% | 12% | 10% | 9% |
| Brand B | | | | | | |
| | Region 1 | 20% | 19% | 18% | 17% | 14% |
| | Region 2 | 5% | 6% | 7% | 6% | 8% |

* Which in turns gets translated into a market share growth at the expense of **Brand B**. On the other hand the declined attention/ promo in
**Region 2** result in a market share loss for **Brand A**.

* Dynamically tracking changes in consumer preferences without waiting for sales results.

* -> Typically changes in consumer behaviour manifest themselves to the brands when the latter analyse the reasons for shifting and/or declining sales in one or many regions. With **Dynamic Portfolio (TM)** brands have an opportunity to observe consumer preferences live, adopt product/pricing/ distribution.

* Measure causal impact of various marketing activities:

* -> Marketing activities typically have a specific objectives vis-a-vis target consumer, competition, share of wallet, etc. Measuring of such activities can only be done either indirectly or in a very crude manner. With **Dynamic Portfolio (TM)** you can observe the impact of your marketing activities virtually in real time.

* Measure whether your promo:
* Is targeted to the right consumer
* Which is competitor is being affected the most
* See competitors counter-moves

## 4.4 Case 5: Engagement and Sales

/// _Given that we are going to be challenged about the size of the consumer base no matter the rational merits of such a claim, I reckon offering our rebate capability as a " rapid testing and experimentation" service might help sell it to bigger brands_

* We offer a **fast deploy, fast feedback** engagement access to out consumer base. You can leverage our extremely precise targeting capabilities and fast consumer response in order to test and tailor-make the main campaign or boost up your sales in its own right....

* Offering consumer rebates if they purchase your products offline and prove the purchase by scanning receipts

* Placing Special QR codes onto products and offering rebates for those who scan at home - users can check their individual rebates in store but claim only at home. Super targeted promo

* Inform users of the ongoing sales in the retailers they frequent..

## 4.5 [Experimental Feature]
**Magic Bucket** - **"I am feeling lucky"** - "I am feeling lucky" type of analysis where a machine learning engine will trying to identify interesting patterns in the data and calculate potential for improvement

1) **Use intelligent clustering to identify hidden patters in your customer base**: