1

SenseBox MIT Sloan G-Lab 2016

Gabriel Awad, Nikolaos Legmpelos, Julia Hawley, Maher Shaban Bogota, Colombia

February 16th, 2017

Young&party Lookingforastyle Matureprofessionals

• 25%ofcustomers• Highlike/mediumexp.

• Formalstyle• Highincome

• 20%ofcustomers• Mediumlike/lowexp.

• Naturalstyle• lowincome

• 55%ofcustomers• Highlike/expertise

• Sexystyle• lowincome

• Youngwomenwhoenjoyusingbeauty&cosmeticsproducts.

• Theyareveryknowledgeabouttheproductsandliketolooksexy.

• Theymaynothavemuchmoneylayingaround,butarewillingtospendonproductsthatmeettheirneeds

• Youngprofessionalswhoareredefiningtheiridentitythroughtheirlooks

• Theyarenotveryknowledgeablewithbeauty&cosmeticsproductsthatfittheir

• Theyhavesignificantmoneytospendonproductsandarelookingtofindabalancebetweensexyandformal.

• Olderwomenwhoenjoyusingbeauty&cosmeticsproducts.

• Theyaresomewhatknowledgeableabouttheproductsandpreferamoreprofessionallook

• Theyhaveplentyofmoneytospendonproductsthatmatchtheirtaste.

~28yrs.~23yrs. ~32yrs.

GeneralassumptionsCostofcapital:y%Grossmargin:x%

DataprovidedbyFernando

Costofcustomeracquisition:~zk COPAveragevaluefromthedataprovided

byFernando

Case-specificassumptions

Customerchurn(afteradjustments,before%ofpayingusers)Datafromlastsection

CostsharingcaseDecreasegrossmargintoS %

intheconsolidatedcase

A B C

FE%ifpayingusers

Datafromlastsection

G

Revenueperbox:24kCOPD

Results Math

Cashflowperperiod: Amountofcashexpectedperconsumerinagivenperiod

NPV: Profitpercustomerafteradjustingforthecostofcapital

Money-over-money:expectedprofitpercustomer(asamultipleofCOCAinvestment)

Internalreturnrate(IRR):expectedreturnofmarketinginvestment

𝐶𝑎𝑠ℎ 𝑓𝑙𝑜𝑤 𝑝𝑒𝑟 𝑝𝑒𝑟𝑖𝑜𝑑 𝒊 = 𝑫 ∗ 𝑨 ∗ 𝑮 ∗ 𝑬𝒊6𝟏 = 𝑯

𝑁𝑃𝑉 = −𝑪 + 𝑁𝑒𝑡 𝑝𝑟𝑒𝑠𝑒𝑛𝑡 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑯

𝑀𝑜𝑛𝑒𝑦_𝑜𝑣𝑒𝑟_𝑀𝑜𝑛𝑒𝑦 = 𝑆𝑢𝑚(𝑯)/𝑪

𝐼𝑅𝑅 = 𝐼𝑛𝑡𝑒𝑟𝑛𝑎𝑙 𝑟𝑒𝑡𝑢𝑟𝑛 𝑟𝑎𝑡𝑒 𝑜𝑓 (−𝑪 + 𝑯 𝑝𝑒𝑟 𝑝𝑒𝑟𝑖𝑜𝑑)

3-clusteranalysis

3clusters Before After Increase Before After Increase

Young-highlike-highexpert 29% 100% 3.4x 3% 100% 36.2xMiddle-lowlike-lowexpert 18% 100% 5.5x 2% 100% 43.7xOld-highlike-middleexpert 23% 100% 4.3x 0% 100%

BrandAwareness BrandTrial

3clusters Before After Increase Before After Increase

Young-highlike-highexpert 46% 100% 2.2x 6% 100% 16.1xMiddle-lowlike-lowexpert 40% 100% 2.5x 2% 100% 40.3xOld-highlike-middleexpert 37% 100% 2.7x 2% 100% 41.0x

ProductAwareness ProductTrial

PurchaseIntent3clusters Promotors Detractors NPS %ofCustomers Yes No Increaseinsales

Young-highlike-highexpert 43% 22% 21 48% 3% 52%Middle-lowlike-lowexpert 48% 20% 27 45% 4% 55%Old-highlike-middleexpert 38% 34% 4 40% 4% 60%

Satisfaction SalesLift

Findings

a) Lowgeneralbrandawarenessandmiddle/highproductawarenessMayindicatethatproductsdonotvaluethebrandandbuytheproductforitsprice

b) MiddleclusterhastheleastawarenessbutthehighestratioofpromotorsandNPSTargetadvertisementstrategiesshouldbemoreprofitableiffocusedonthispersona/cluster

c) OldclusterhasthehighestnumberofdetractorsandlowestNPSTargetadvertisementstrategiesshouldavoidthispersona/cluster

Photos: SenseBoxOffice

Cartagena

BogotaGolfClub

Chia

Analytics report for suppliers (1/3) - Overview

Howclusterizationofdemographicdatacanincreasemakedatavisualizationeasier/moreinsightful

Currentpresentation Dataclustering

• Hard-to-interpretinformation• Highlycorrelateddatapresentedseparately

Twowaystogroupthedata

Subjectivecriteriamanagementdefineshowtoclusterthedata(e.g.,urbanheavyusersofmake-up)

ObjectivecriteriaStatisticaltoolsdefinehowtoclusterthedata(e.g.,urbanheavyusersofmake-up)

• Intuitivevs.data-drivenapproach

Output

Urbanheavyusersofmake-up(35%)

Topclusters(example)

Low/mediumuserswithformal/classicstyle(30%)

Young,party/sexystyleusers(20%)

• SmallernumberofgroupswillproviderbetterunderstandingofSensebox’s customerbase

Others(15%)

Findings

Performancewithingroup

Awareness Intent

80% 30%

20% 90%

60% 50%

60% 60%

• Group1:Lowreturnonmarketinginvestment• Group2:Highreturnonmarketinginvestment

1

2

3

4

Analytics report for suppliers (2/3) - Personas

Analytics report for suppliers (3/3) - Metrics

Clusterization of demographic and psychographic data enables SenseBox to improve their understanding of their client base

SenseBox has 3 types of customers with varying ages, make-up expertise, styles and hobbies

Evaluating the product metrics for each persona/cluster improves SenseBox’s value proposition to its suppliers

Calculating the Customer LTV will help SenseBox optimize its marketing and investment strategies

Financials - Customer lifetime value

IT systems

IT systems applicable to SenseBox’s business model were recommended; a CRM and ERP platform were set up.

Insights with Google Analytics