This document provides a summary of insights from analyzing transaction data across different regions for a food delivery app. Key findings include:
1. Consumption patterns vary significantly between countries like Finland and Poland, with higher usage on weekends in both.
2. New user registration dips on Thursdays but peaks on Fridays and Sundays, matching consumption patterns.
3. Fraud levels differ in regions, with above average fraud in EUR, GEL, and SEK currencies. Sweden sees fraud primarily through Mastercard transactions.
4. Features like split payments and happy hour tokens are recommended to boost usage during low periods based on regional behavior analysis.
2. Table of Contents
WORK PACKAGE 1: ACROSS GEOGRAPHIC REGION ....................................................................................................... 3
Study 1: All events across Geographic Region.............................................................................................. 3
Study 2: Consumption Pattern across regions (Finland and Poland)............................................................ 4
New Feature: Happy Hour Token ................................................................................................................. 9
New Feature: Split (Payment)....................................................................................................................... 9
Study 3: New User Registration Pattern across regions ............................................................................. 10
WORK PACKAGE 2: SWEDEN .................................................................................................................................. 11
Study 1: Fraud in Sweden........................................................................................................................... 11
WORK PACKAGE 3: POLAND .................................................................................................................................. 16
Study 1: Events in Poland ........................................................................................................................... 16
Study 2: Relative Transaction in Poland ..................................................................................................... 17
Study 3: Consumer Behaviour in Digital Payments .................................................................................... 20
Study 4: Digital Payment Adoption in Poland............................................................................................. 22
INSIGHTS THAT WERE NOT PLOTTED AND DISCUSSED .................................................................................................... 26
PART C......................................................................................................................................................... 27
Feature 1: Split (recapitulating).................................................................................................................. 27
Feature 2: Wallet........................................................................................................................................ 27
Feature 3: Wolt Combine ........................................................................................................................... 28
WORK PACKAGE 4: WALLET................................................................................................................................... 31
User Persona .............................................................................................................................................. 31
User Story Template................................................................................................................................... 31
Wallet Product Requirement Document (PRD).......................................................................................... 32
Wallet Development Flow.......................................................................................................................... 32
Development Effort Cost............................................................................................................................ 33
Scrum Method for Managing Product Backlog........................................................................................... 34
REFERENCES:....................................................................................................................................................... 35
3. Work Package 1: Across Geographic Region
Study 1: All events across Geographic Region
Figure 1: All Events (€)
Event Code Currency
0K 500K 1000K 1500K 2000K 2500K 3000K 3500K
Value in Euro
AUTHORISATION CZK
DKK
EUR
GEL
HUF
ILS
NOK
PLN
SEK
CANCELLATION CZK
DKK
EUR
GEL
HUF
NOK
PLN
SEK
CAPTURE CZK
DKK
EUR
GEL
HUF
NOK
PLN
SEK
CHARGEBACK EUR
SEK
USD
NOTIFICATION_OF_CHAR.. EUR
SEK
USD
NOTIFICATION_OF_FRAUD EUR
GEL
SEK
REFUND CZK
DKK
EUR
GEL
NOK
PLN
SEK
REQUEST_FOR_INFORMA.. EUR
1,078.64K€
3,261.25K€
348.92K€
103.06K€
140.14K€
60.24K€
57.56K€
13.52K€
0.00K€
13.41K€
30.52K€
88.56K€
1.65K€
1.14K€
3.67K€
0.52K€
2.47K€
2,968.03K€
307.63K€
985.77K€
124.80K€
55.77K€
50.46K€
11.53K€
85.21K€
8.41K€
5.10K€
1.01K€
6.67K€
5.10K€
1.01K€
8.43K€
0.06K€
6.90K€
0.03K€
9.09K€
1.41K€
0.06K€
0.48K€
0.01K€
1.00K€
0.19K€
All Events (€)
Sum of Value in Euro for each Currency broken down by Event Code. The marks are labeled by sum of Value in Euro. The view is filtered on Event Code,
which excludes event_code.
4. Few Key Insights:
Considering zones in EUR (Finland, Latvia, Lithuania in this dataset)
- Ideally, all authorisation should be able to capture. But in many cases, that is not happening
as it is entering the states: Cancellation, Chargeback, Fraud or Refund primarily. For example,
when we consider business value of EUR Zone i.e. Finland, Latvia and Lithuania (in this dataset)
– Authorisation has been of 3261.25K € but capture has been only 2968.03K €. A loss of
information of business 293.22K €.
- Chargeback of 8.41K € but notification of only 6.67K €. Loss of amount of 1.74K €.
Few reasons can be - either Customer Success Teams has not able to grow steadily with
exponential rise in business, peak time bounce off or not able to trace notification of
chargeback, henceforth could be a form of fraudulent behaviour.
- Chargeback in other regions like SEK (Sweden) and USD (USA) is excellent. Every Chargeback
request has a follow up of Chargeback Notification; except EUR Region (Finland, Latvia,
Lithuania in this dataset)
In Denmark (DKK):
- Authorisation has been of 1078.64k €, whereas capture is 985.77K €. The other two states of
the payment events are: Cancellation and Refund.
- Cancellation is of 30.52K € and Refund is of 9.09K €, which means 3.67% of transaction in 2
weeks has been called off by the customer. It might be reflecting on the customer satisfaction
in Denmark Region. The event states of cancellation and refund might be reflecting on the
kind of restaurant and service providers Wolt has tied up with in Denmark.
Fraudulent Behaviour:
- Fraudulent behaviour is primarily in the region of EUR (Finland, Latvia and Lithuania), GEL
(Georgia) and SEK (Sweden).
Study 2: Consumption Pattern across regions (Finland and Poland)
When I plotted the consumption pattern across geographic region, it showed wide variation in each
geographic zone. For which, I needed to plot country-wise (Merchant Code). For example,
consumption pattern across the week varied widely between Finland, Sweden and Poland.
But for the brevity of the assignment, I will focus on Case 1: Finland, because I am aware of the social
construct a bit more and Case 2: Poland, as the business is weak, it was interesting to study further.
I am adding the calendar of October 2018, for reference.
5. CASE 1: Finland
Figure 2 : Consumption pattern across week – in Finland
Few Key Insights – Consumption in Finland:
- The consumption drops on Thursdays before it takes a peak rise on Fridays. The reason might
be, people are saving up their “eating out quota” to eat out over the weekend.
- On Fridays, it has the second highest peak, possibly because people are ordering food to
celebrate TGIF with colleagues. The timing might be during afternoon lunch or as evening food
and snacks to go with drinks.
- On Sundays, it has the highest peaks. Assumption is, with understanding Finland’s social
context, people love spending their leisurely Sundays with family. They might be ordering in
food to eat with family. This will usually be the afternoon lunch or dinner.
Let’s try to drill down further and check the consumption pattern across hours on Fridays and Sundays.
Merchant Account Code
Wolt_FIN
1 Oct 2 Oct 3 Oct 4 Oct 5 Oct 6 Oct 7 Oct 8 Oct 9 Oct 10 Oct 11 Oct 12 Oct 13 Oct 14 Oct
Day of Event Date [October 2018]
0K
100K
200K
300K
400K
Value in Euro
Consumption Pattern across Week
The trend of sum of Value in Euro for Event Date Day broken down by Merchant Account Code. The view is filtered on Merchant Account Code, which keeps
Wolt_FIN.
6. Figure 3 : Consumption in Finland on a Friday
Figure 4 : Consumption in Finland on a Sunday
Few Key Insights:
- On both Fridays (5th
and 12th
October 2018), the peak consumption over the day has happened
over 16:00 hour i.e. close to the end of office hours. We can assume people are getting ready
to celebrate TGIF (“Thank God it’s Friday”) with colleagues. They might be ordering in snacks
and finger food to go with their drinks. We could have analysed further if we could have got
more details on the restaurants and food type, they have placed orders for.
7. - On both Sundays (7th
and 14th
October 2018), the peak consumption over the days has
happened at 14:00-15:00. It definitely matches with our assumption of families wanting to
have a leisurely lunch together. Finland usually has their lunch between 11:30-13:00. We can
assume, probably, the lunch ordered might be something light and healthy which will be
consumed after a late morning (heavy and hearty) breakfast. We could have said with assured
if we had more data on the restaurants and food type, they have placed orders for.
CASE 2: Poland
Figure 5: Consumption pattern across week – in Poland
Few Key Insights:
- In Poland, the consumption peak is on Fridays (5th
and 12th
October 2018) and Sundays (7th
and 14th
October 2018). It might be following the same pattern as of Finland- Friday high peak
consumption is because of TGIF and Sunday high peak consumption is because people might
be celebrating with family and friends.
- There are two major dip points in Poland. First, during early weekdays like Tuesdays (2nd
October 2018) and Wednesdays (10th
October 2018). Second one prominently lies between
the highest peak points – which is Saturday.
If we drill down further, we can understand what kind of consumption happens over Friday and
Sundays.
Merchant Account Code
Wolt_POL
1 Oct 2 Oct 3 Oct 4 Oct 5 Oct 6 Oct 7 Oct 8 Oct 9 Oct 10 Oct 11 Oct 12 Oct 13 Oct 14 Oct
Day of Event Date [October 2018]
0
500
1000
1500
2000
2500
Value in Euro
Consumption Pattern-week (POL)
The trend of sum of Value in Euro for Event Date Day broken down by Merchant Account Code. The view is filtered on Merchant Account Code, which keeps
Wolt_POL.
8. Figure 6: Consumption in Poland on a Friday, 13:00 Hour
Figure 7: Consumption in Poland on a Friday, 16:00 Hour
Key Insight:
Poland, across all days usually have two distinct consumption peak – at 13:00 hour and 16:00 Hour.
The consumption can be because of ordering lunch and snacks with drinks, respectively.
Recommendations from Consumption Pattern Study:
9. Each Geographic zone has different consumption patterns. This information can be valuable to
Marketing & Sales and Customer Success Teams.
- We can have advertisement tailored for each geographic zone, based on their cultural context,
norms and desires.
- Advertisements on social media can be running for specific days and hours, either for dip days
or peak days. We might have to do a/b testing to understand which brings more business
growth.
- Moreover, we can run campaigns (goodies, freebies, tokens and credits) targeting specific
trends.
New Feature: Happy Hour Token
1. We can introduce Wolt Happy Hour Token, different than a normal Token. These tokens can
be used during dip days and hours (like Tues – Thurs, between 13:00-16:00)
2. These tokens can be bought in bulk (a pack of 3, 5,7) and used only during happy hours (varies
with each geographic region).
3. During payments, when we add the Happy Hour Token, Delivery Charges goes flat reduced by
certain % (like 25-40%).
New Feature: Split (Payment)
Consumer Behaviour Understanding:
We have seen the spike in consumption on Fridays and Sundays. On Fridays, the consumption style is
more towards the evening hours (16:00-18:00), when one spends good time with their friends and
colleagues. This is a moment, where one person orders for the whole group.
Leveraging this Consumer Behaviour: Split Payment
When a user is ordering for a group, during payment, he (/she) can split the payment, by adding
friend’s Accounts. Result is, total bill splitted and charged from the individual accounts.
This way, their accumulative delivery charge is reduced, and user experience enhanced. It also re-
enforces the idea of “account” and encouraged to be used for collecting “Wallet” (to be introduced
later) Tokens and Credits.
Figure 8 : Split Feature
10. Study 3: New User Registration Pattern across regions
Figure 9: New Registrations across regions
Few Key Insights:
- On an average, there is a dip across geographic regions on Thursdays (4th
and 11th
Oct 2018).
- There is a peak in new user registration on Fridays (5th
and 12th
Oct 2018) and Sundays (7th
and 14th
Oct 2018).
- Overall, there is a steady growth in new user registration steadily over the week from Monday
onwards till Sunday.
Recommendations from New User Registration study
This information can be valuable to Marketing & Sales and Customer Success Teams.
- With the already existing concepts of Token and Credits, we can have targeted campaigns on
Thursdays. This will be good for the dip and for the rise in new user registration on following
days.
- Splitwise can drive more new user registrations onto the Food Delivery App.
1 Oct 2 Oct 3 Oct 4 Oct 5 Oct 6 Oct 7 Oct 8 Oct 9 Oct 10 Oct 11 Oct 12 Oct 13 Oct 14 Oct
Day of Event Date [October 2018]
0K
2K
4K
6K
8K
10K
12K
14K
16K
18K
Distinct count of Customer Id
New Registrations across Regions
The trend of distinct count of Customer Id for Event Date Day.
11. Work Package 2: Sweden
Study 1: Fraud in Sweden
SEK (Sweden) was one region, which was interesting to study further as it had pretty much all the
payment event states, namely: Authorisation, Cancellation, Capture, Chargeback, Notification of
Chargeback, Notification of Fraud and Refund.
All Events in Sweden across payment methods (Amex, Amex-Applepay, Mastercard, Visa, etc)
Figure 10: Events in Sweden
Currency / Event Code
SEK
AUTHORISATI.. CANCELLATION CAPTURE CHARGEBACK NOTIFICATION..NOTIFICATION.. REFUND
0K
50K
100K
150K
200K
250K
300K
350K
Value in Euro
348.92K€
307.63K€
13.41K€
5.10K€ 5.10K€ 6.90K€
1.00K€
Events in Sweden
Event Code
AUTHORISATION
CANCELLATION
CAPTURE
CHARGEBACK
NOTIFICATION_OF_CHARGEBACK
NOTIFICATION_OF_FRAUD
REFUND
Sum of Value in Euro for each Event Code broken down by Currency. Color shows details about Event Code. The marks are
labeled by sum of Value in Euro. The view is filtered on Currency and Event Code. The Currency filter keeps SEK. The Event
Code filter excludes event_code and REQUEST_FOR_INFORMATION.
12. All Events in Sweden through Mastercard
Figure 11: Sweden-MC Fraud
Few Key Insights for Sweden (one of my Case Country):
- Fraudulent behaviour in Sweden across all payment methods has been of 6.90K €. When we
check payment behaviour through Mastercard – Fraudulent behaviour is 6.90K €. This shows,
that all fraudulent behaviour in Sweden has happened through Mastercard.
- In Sweden, across all payment method, Chargeback is of net 5.10K €. When we dig deeper and
try to analyse, the behaviour through Mastercard, the net Chargeback is same – net 5.10K €.
This shows, that all chargeback behaviour in Sweden has happened through Mastercard. This
is same as Fraudulent behaviour. We can assume, that chargeback request in Sweden is a form
of Fraudulent behaviour.
Note : At this stage, I wish to look into how fraudulent behaviour in card payments can be reduced
overall, especially because of the special case of high fraudulent behaviour in Sweden.
Study [5] by European Central Bank shows that there is a correlation between countries with high
usage of bank cards and instances of fraud occurrence. Thus, when our society becomes more digital
and adapted to digital payments methods more eloquently, there will be spike in fraudulent behaviour
in payments.
Research 1: Fraud Detection and SD2 Strong Customer Authentication (SCA) Directive
The dataset provided is of 2018, since then in September 2019, there already has been SD2 Strong
Customer Authentication (SCA) directive in place to protect from fraudulent behaviour in payments
industry [17]. This led to enforcement of strong customer authentication mandatory for all; payments
done through online medium within European Union, which means every online purchase above €30
and every 6th
purchase, will require the “3D authentication”.
Currency / Payment Method / Event Code
SEK
mc
AUTHORISATI.. CANCELLATION CAPTURE CHARGEBACK NOTIFICATION..NOTIFICATION.. REFUND
0K
50K
100K
150K
200K
250K
Value in Euro
241.90K€
216.09K€
8.73K€ 5.10K€ 5.10K€ 6.90K€
0.57K€
Sweden-MC Fraud
Event Code
AUTHORISATION
CANCELLATION
CAPTURE
CHARGEBACK
NOTIFICATION_OF_CHARGEBACK
NOTIFICATION_OF_FRAUD
REFUND
Sum of Value in Euro for each Event Code broken down by Currency and Payment Method. Color shows details about Event
Code. The marks are labeled by sum of Value in Euro. The view is filtered on Currency, Event Code and Payment Method. The
Currency filter keeps SEK. The Event Code filter excludes event_code and REQUEST_FOR_INFORMATION. The Payment
Method filter keeps mc.
13. Research 2: 3D Authentication and User Experience Friction
3D Secure is also known as a payer authentication, which is a security protocol preventing from
fraudulent behaviour in online debit and credit card transaction [19]. This initial security payment
method was created by Visa and Mastercard and is known as “Verified by Visa” and “MasterCard
SecureCode” respectively.
This makes sense for the two big giants to come together and form this payment authentication step.
From the dataset, we saw that two most used payment methods were of Mastercard and Visa.
Figure 12 : Most used Payments Method - Mastercard and Visa
What 3D secure authentication does is, creates and assigns a password (which needs to be verified)
every time a user is making an online transaction [18]. It is a called a 3D secure Pay because of 3 parties
being involved are:
- The Company the purchase is being made from
- The acquiring bank (bank of the company)
- And VISA and MasterCard
But security is still a problem when the user is not:
- In European Union like users in Georgia and Israel
- Or a user is not using Mastercard or Visa, maybe they are using AMEX.
Payment Method
Null amex amex_ap.. mc payment.. unknown.. visa
0K
20K
40K
60K
80K
100K
120K
140K
160K
180K
200K
Number of Records
Most Payment Method
Sum of Number of Records for each Payment Method.
14. Figure 13 : Activity in AMEX Payment Method
Insight:
Although, when I plotted Amex card activity types across all geographic places, there has been no
specific instance of fraudulent behaviour, but we cannot be 100% assured that Refund and Request
for cancellation is not a form of fraudulent behaviour.
Figure 14 : Fraudulent behaviour in Georgia
Insight:
In Georgia, of total 367€ was of fraud in just 2 weeks. We cannot just skip this and rely on the SD2
Strong Customer Authentication (SCA) directive, which is in place only in European Union.
There can be ways we can add security measures to Wolt Payment Gateway.
Payment Method / Event Code
amex mc visa
REFUND REQUEST_FOR.. NOTIFICATION_.. REFUND NOTIFICATION_.. REFUND
0K
1K
2K
3K
4K
5K
6K
7K
8K
9K
Value in Euro
276€ 186€
8,696€
3,890€
6,706€
7,917€
AMEX Fraud
Sum of Value in Euro for each Event Code broken down by Payment Method. The marks are labeled by sum of Value
in Euro. The view is filtered on Payment Method and Event Code. The Payment Method filter keeps amex, mc and
visa. The Event Code filter keeps event_code, NOTIFICATION_OF_FRAUD, REFUND and
REQUEST_FOR_INFORMATION.
Merchant A..
Event Code
AUTHORISATION CANCELLATION NOTIFICATION_OF_FRAUD REFUND
0K 50K 100K 150K
Value in Euro
0K 50K 100K 150K
Value in Euro
0K 50K 100K 150K
Value in Euro
0K 50K 100K 150K
Value in Euro
Wolt_GEO 103,075€ 2,466€ 367€ 61€
Fraud in Georgia
Sum of Value in Euro for each Merchant Account Code broken down by Event Code. The marks are labeled by sum of Value in Euro. The view is filtered on
Event Code and Merchant Account Code. The Event Code filter keeps AUTHORISATION, CANCELLATION, NOTIFICATION_OF_FRAUD and REFUND. The
Merchant Account Code filter keeps Wolt_GEO.
15. Research 3: Integrating a security mindset
Managing fraud shouldn’t be an addon but a holistic way of allowing more security in itself, which
forms trust for our users. Moreover, as the Food Delivery App is growing in different continents and
falls under different governing zones, it has become increasingly important to integrate a safe
payment gateway for our users specially outside European Union like Israel, Georgia, USA, etc.
Ways to integrate Security:
Following can be the ways to integrate authentication [15]:
1. Two Factor Authentication (2FA), such as confirming texts, codes, calls after a certain safe bill
limit
2. Tokenization [16]: the customer’s primary account number (PAN) will be replaced by a series
of randomly generated numbers called as “Tokens”, which is safer to pass through series of
wireless networks. This helps in not exposing the actual bank details. Although, it is decade
old mechanism, it is compliant with Payment Card Industry (PCI) and works with existing POS
systems [15]. It is more cost-effective than Encryption.
Figure 15 : Tokenization Simplified [16]
3. Encryption [15][16]: Encryption transforms bank details data into an unreadable form, even
to those who can access the encrypted data. Encryption has lot of advantages like cloaking
private messages in P2P apps, transferring sensitive information in vulnerable environment,
etc. But the method of Encryption in costly.
Suggestion: Tokenization method to be adopted as a security measure with Wolt Payment
gateway.
1. In the Payments industry, there is a shift back to the Tokenization method over Encryption.
Few reasons [15][16] for it are: Tokenization is more Cost-effective, secure and reduces the
scope of PCI compliance further over Encryption.
2. In the second of this document, I introduce a concept of Wolt Wallet – in this kind of feature,
Tokenization will allow users to store credit card information in Wolt Wallet without exposing
the original card information.
16. Work Package 3: Poland
Study 1: Events in Poland
When I studied, the card transaction behaviour in Poland, it did not have a wide variety of event types
like Sweden. One, understanding can be that the Food Delivery App in Poland is quite new, and fraud
and corruption creeping into business is less.
For example – I studied the transaction behaviour across all payment methods and then tried
understanding cancellation behaviour more. My objective was to trace if the behaviour in Poland is
similar to Sweden – that all Cancellation (or dubious behaviour) will be from a particular kind of
payment method.
All Transaction type across all payment method
Figure 16: Events in Poland
Currency / Event Code
PLN
AUTHORISATION CANCELLATION CAPTURE REFUND
0K
2K
4K
6K
8K
10K
12K
14K
Value in Euro
13.52K€
11.53K€
0.52K€
0.01K€
Events in POL
Event Code
AUTHORISATION
CANCELLATION
CAPTURE
REFUND
Sum of Value in Euro for each Event Code broken down by Currency. Color shows details about Event
Code. The marks are labeled by sum of Value in Euro. The view is filtered on Currency, which keeps PLN.
17. Figure 17: POL-Cancel
Few Key Insights:
- All cancellation happened primarily through Mastercard and Visa. Total business worth of
0.52K € has been cancelled.
- Total business in Poland is one of the lowest.
Note:
Through these 2 graphs I could not delve much interesting insights. I decided to then study Poland in
a relative term. Is the business in Poland one of the lowest, how popular is the Food Delivery App in
Poland, how many unique and repetitive customer in Poland, etc.
Study 2: Relative Transaction in Poland
Figure 18: Net Value in each Region
Currency / Event Code / Payment Method
PLN
CANCELLATION
mc visa
0
50
100
150
200
250
300
Value in Euro
0.21K€
0.31K€
POL-Cancel
Sum of Value in Euro for each Payment Method broken down by
Currency and Event Code. The marks are labeled by sum of
Value in Euro. The view is filtered on Currency and Event Code.
The Currency filter keeps PLN. The Event Code filter keeps
CANCELLATION.
Currency
currency
CZK
DKK
EUR
GEL
HUF
ILS
NOK
PLN
SEK
USD 2,018€
688,055€
25,581€
269,087€
5€
117,152€
190,862€
6,342,957€
2,104,018€
109,701€
€
Net Value in each Region
Sum of Value in Euro broken down by Currency.
18. Few Key Insights:
- Poland has the second lowest business, while ILS Region or Israel has the lowest.
- The only plausible reason for Israel having a net business of 5€ (in 2 weeks) is maybe Wolt is
at its earliest business touchpoint in Israel.
Question: Maybe, the Food Delivery App has just entered Poland just recently (as when the data was
collected).
To test my assumption, I will try and check how many transactions has been recorded, unique and
repetitive customers.
If percentage of unique customers is high and total customers are quiet low, we can assume that Wolt
has just entered Poland recently (at the time of data collection).
Figure 19: Unique and Total Customers per Geographic Area (Finland Pointed out)
Focussing and Zooming into Poland
Figure 20: Poland - Unique and Total Customers
19. Few Key Insights:
- Out of 3685 total customer ID, 1373 are new and unique customer ID. A whopping 37.26% are
new customers i.e. 1/3 of the business is new business.
- Moreover, total transaction unit is 3685 – this reflects that business in Poland is growing and
new.
Figure 21: Hungary - Unique and Total Customers
Figure 22: Czech Republic - Unique and Total Customers
Few Key Insights:
- In Hungary, percentage of unique customer is 31.22%.
- In Czech Republic, percentage of unique customer is 29.22%.
- In all 3 cases of Poland, Hungary and Czech Republic – percentage of unique (distinct)
customers reflect nearly 30% of the total business. These regions will be vital for Wolt’s
growth.
- Also, these regions might share similar consumer behaviour related to digital payments and
trust.
20. Geographic Region % Distinct Customer Entered
Czech Republic 29.22 First
Hungary 31.22 Second
Poland 37.26 Third
Key Insight from this gut feeling:
- The Food Delivery App is gradually over time reaching a cap, in growing digital economy zones
like Czech Republic, Hungary and Poland. As its operation grow older, decline of adoption of
the app by new and distinct customer.
- The cap or the limit is a constraint put by a certain kind of consumer.
Question is: Who is this consumer – what kind of consumer behaviour they reflect?
Study 3: Consumer Behaviour in Digital Payments
To study the above question of Consumer Behaviour, I wanted to look into adoption of digital
payments. I plotted a comparative study on Card transaction in Sweden, Estonia, Czech Republic,
Hungary and Poland.
Figure 23: Comparative study of Card Transaction in Poland and Sweden
Few Key Insights:
According to Wikipedia Timeline [1], the Food Delivery App was established in Sweden then Estonia
in 2016. And, in 2018, serially in Czech Republic, Hungary and Poland.
It is but understandable, that Sweden and Estonia will have higher number of transaction than Czech
Republic, Hungary, and Poland. Also, Sweden should have higher number of transactions than Estonia,
as it was established earlier than in Estonia. But when we look between Sweden and Estonia, Estonia
has way higher number of transaction (almost double).
Payment M..
Merchant Account Code
Wolt_SWE Wolt_EST Wolt_CZE Wolt_HUN Wolt_POL
0K 20K 40K 60K 80K
Number of Records
0K 20K 40K 60K 80K
Number of Records
0K 20K 40K 60K 80K
Number of Records
0K 20K 40K 60K 80K
Number of Records
0K 20K 40K 60K 80K
Number of Records
amex
mc
visa
489.17K€
162.41K€
41.72K€
958.19K€
151.62K€
6.97K€
72.86K€
35.06K€
1.78K€
88.32K€
27.65K€
1.18K€
11.54K€
14.02K€
0.01K€
Comparative Card Transaction
Sum of Number of Records for each Payment Method broken down by Merchant Account Code. The marks are labeled by sum of Value in Euro. The view
is filtered on Merchant Account Code, which keeps Wolt_CZE, Wolt_EST, Wolt_HUN, Wolt_POL and Wolt_SWE.
21. - Premise 1: Maybe, number of transactions in each geographic region is not only dependant
on time of establishment of Wolt operation but maybe on other factors like adoption of digital
payment, unrestricted access to online services, penetration of internet connection in the
country, etc.
- Premise 2: It also depends on the competitive landscape, tie ups with restaurants and options
available to customers.
- Premise 3: Customer Service, Chargeback request followed through, Refund, etc – which
determines happy and satisfied return customers.
- Premise 4: The competitive landscape can be fierce.
For brevity of the assignment, I will work with Premise 1.
Understanding Premise 1: Adoption of Digital Platforms and Payment
Figure 24 : Estonia - best country for Digitally Connected Life [3]
Estonia is being considered the best country and ranks 1st out of 68 countries when it comes to the
best and worst countries to live a connected life, by a 2019 Internations Report. This reflects on how
Estonians have quickly adapted digital way of life, trust on digital platform and reliance on a digitally
connected life. This might also be reflected in their usage of digital app like Wolt.
Estonia has almost double the usage adoption than in Sweden (at the time of data collection).
Inference:
This might also give us a hint, why regions like Czech Republic, Hungary and can reach a cap quicker in
terms of new growth and newer customer acquisition. Consumers might not be used to relying and
having trust on digital platforms and payments.
We have to look into digital payment adoption pattern in these regions. For the case of assignment, I
will look into Digital Payment Adoption in Poland, one of my case regions for the assignment.
22. Study 4: Digital Payment Adoption in Poland
Study [5] by European Central Bank shows that there is a correlation between countries with high
usage of bank cards and instances of fraud occurrence. In the Figure 28 below, we see that countries
like Denmark, Sweden and Estonia has high Fraud occurrence. In Poland, Fraud occurrence has
not happened yet.
Figure 25 : Fraud in Region - Sweden, Denmark, Estonia and Poland
According to Figure 27 “Estonia - best country for Digitally Connected Life” – Denmark, Sweden and
Estonia are in top 5 countries for being most digitally connected, which means adoption of digital
platforms, digital payments and card payments are higher.
From Figure 27 and Figure 28, we can infer and assume that Poland has lower penetration of bank
accounts and adoption of payment through cards.
Moreover, I believe that adoption of payment through cards is just a tip of the iceberg. There are lot
of social and technological issues which will result in adoption of payments through cards, henceforth
usage of Wolt (as a part of it). For example, Oxley and Yeung (2001) mentions that e-commerce and
digital platform adoption depends upon many factors like availability of the internet, customer’s trust
in online transactions, account balance readiness and household’s disposable income per capita.
I would like to focus and research more on few of the factors like:
1. Household’s disposable income
2. Card Payment as relative to other payment methods.
Merchant Account Code Event Code
0K 100K 200K 300K 400K 500K 600K 700K 800K 900K 1000K 1100K 1200K
Value in Euro
Wolt_SWE REFUND
CHARGEBACK
NOTIFICATION_OF_CHAR..
NOTIFICATION_OF_FRAUD
CANCELLATION
CAPTURE
AUTHORISATION
Wolt_DNK REFUND
CHARGEBACK
NOTIFICATION_OF_CHAR..
NOTIFICATION_OF_FRAUD
CANCELLATION
CAPTURE
AUTHORISATION
Wolt_EST REFUND
CHARGEBACK
NOTIFICATION_OF_CHAR..
NOTIFICATION_OF_FRAUD
CANCELLATION
CAPTURE
AUTHORISATION
Wolt_POL REFUND
CANCELLATION
CAPTURE
AUTHORISATION
353.71K€
307.63K€
13.41K€
1.00K€
5.32K€
5.32K€
6.90K€
1,078.59K€
985.77K€
30.52K€
9.09K€
1.84K€
3.14K€
2.23K€
578.16K€
516.96K€
17.58K€
0.10K€
1.16K€
1.59K€
1.24K€
11.53K€
13.52K€
0.01K€
0.52K€
Fraud in Regions
Sum of Value in Euro for each Event Code broken down by Merchant Account Code. The marks are labeled by sum of Value in Euro. The view is filtered on
Event Code and Merchant Account Code. The Event Code filter keeps 8 of 8 members. The Merchant Account Code filter keeps merchant_account_code,
Wolt_DNK, Wolt_EST, Wolt_POL and Wolt_SWE.
23. 3. Account Balance Readiness
Research 1: Household’s disposable income
According to JP Morgan 2019 Payments Trend Reports [7], Poland has an average annual online
expenditure per capita as €581. This quiet a low figure in comparison with the rest of Europe
average as €2186. This expenditure figure is linked to the economic development of the country
and its legacy as a former soviet state. According to JP Morgan Report, only Czech Republic and
Poland has an average expenditure per capita below €1000.
This is linked also to the consumer behaviour in Poland being highly cost conscious.
I tried to understand if this information matches with the dataset provided.
Figure 26: Average Ticket Size (Bill Size) across Europe
In Figure 29, I calculated average ticket (bill) size across Europe. Poland was the only country which
had average bill size lower than €10. This matches the information of Poland being an economically
developing country and consumers being highly cost conscious.
Although, the question still remains if the ticket size is way lower because of nation being in the
developing stage. To analyse further, I took the case of Poland and Hungary.
According to countryeconomy.com information [8], in 2018, Hungary had a GDP Per capita as 16,484
$ and Poland has GDP per capita as 15,538 $. Both countries have almost equivalent GDP per capita.
Also, Hungary has an average household expenditure per capita in 2018 as $ 4,526.036 [9] and Poland
has an average household expenditure per capita in 2018 as $ 3,943.366.
The difference in the average household expenditure between Hungary and Poland is by 12.87%. But,
the ticket size difference between Hungary and Poland is almost 40%. There can be many reasons for
this difference, following can be amongst them:
- Poland is highly cost conscious
- Card Payment as less trustworthy, people try and bill lesser on online apps.
Poland being a highly cost-conscious society has been found in research conducted by JP Morgan [7]
and other consulting firms [11]. The adoption rate of digital apps like Wolt can take time as they seek
lot of assurance before making any purchase. Consumers seek to compare prices, ratings, reviews,
Merchant Account Code
0 5 10 15 20 25 30 35 40
Avg Ticket
Wolt_NOR
Wolt_DNK
Wolt_FIN
Wolt_SWE
Wolt_EST
Wolt_LVA
Wolt_LTU
Wolt_CZE
Wolt_HUN
Wolt_POL
merchant_account_code
35.68€
32.82€
26.43€
24.51€
19.42€
19.15€
16.51€
14.03€
11.54€
6.94€
Ticket Size
Avg Ticket for each Merchant Account Code. The marks are labeled by Avg Ticket. The view is filtered on Merchant Account Code, which excludes Wolt_GEO.
24. before committing to make a purchase. Almost 62% uses comparison website before making a
purchase [12].
Few Key Insights (summing up):
- Poland has the lowest ticket size, way lower than Hungary which has almost equivalent
average household expenditure per capita.
- Poland is highly cost-conscious society: Confirmed with various studies.
- Card Payment as less trustworthy, people try and bill lesser on online apps: We have to find
out more.
Research 2: Card Payment as less trustworthy in Poland
Figure 27 : E-Commerce payment method split in Poland [7]
In 2018, according to JP Morgan studies [7], Poland used Banking as primary method of payment,
followed by Bank cards at 22%. But by 2021 Bank Cards will take up 25% of market share and digital
wallet will have higher market share of 30%.
Few Key Insights:
- Card payments are considered less trustworthy – is not completely true.
- Bank Card Payment and Digital Wallet will take up 55% of market share by 2021, which is good
for the Food Delivery App.
- Digital Wallet will eventually overtake Bank Card payments. This means, that more and more
transfer through wallet will reduce share % earned by Wolt.
Can we build a Digital Wallet Services?
Research 3: Account Balance Readiness in Poland
Account Balance Readiness is referred to having sufficient fund in bank account, to make an
immediate purchase. According to a study by PayU [13], an estimated 42% of polish online consumers
lack sufficient funds in their banking accounts to make an immediate purchase. For this reason, PayU
is launching a service, where consumers can have the ability to defer their e-commerce payment upto
30 days.
25. This also, might explain the reason, why Poland has such a low-ticket (bill) size.
Figure 28 : Average Ticket (Bill) Size lowest in Poland
Feature Recommendations:
There can be two of the many things that can be strategized for Payment Service in Poland.
1. As more and more consumers are moving towards Digital Wallet Payment, by 2021 market
share of payments will reach 30%. If the Food Delivery App relies too much upon digital wallet
service providers, then profit% will go down for Wolt.
Wolt needs to build a Wallet Service for its consumers in Poland.
2. Tie-up with more and more Digital Wallet Service providers, especially with PayU for their 30
days pay later service, which is can be used by 42% of the consumers [13].
Figure 29: Top digital service providers in Poland [14]
3. Moreover, as users often don’t have sufficient fund to make an immediate payment, it can be
useful to allow them to pay through combination of different payment methods like Cash on
Delivery (COD), combination of 2 different Bank Cards.
Wolt can add a new feature in Digitally Developing Economies like Poland, Georgia, etc to pay
through Combine. Wolt Combine feature- facilities combination of payment through 2
payment methods at once.
Wallet will be useful in this scenario too, ensuring that people have enough cash in their Wolt
Wallet to make an immediate transaction.
Merchant Account Code
0 5 10 15 20 25 30 35 40
Avg Ticket
Wolt_NOR
Wolt_DNK
Wolt_FIN
Wolt_SWE
Wolt_EST
Wolt_LVA
Wolt_LTU
Wolt_CZE
Wolt_HUN
Wolt_POL
merchant_account_code
35.68€
32.82€
26.43€
24.51€
19.42€
19.15€
16.51€
14.03€
11.54€
6.94€
Ticket Size
Avg Ticket for each Merchant Account Code. The marks are labeled by Avg Ticket. The view is filtered on Merchant Account Code, which excludes Wolt_GEO.
26. Insights that were not plotted and discussed
Following are the insights that I left out of scope:
- Revenue lost through fraudulent behaviour in payment transcations (possible to find out
with a guesstimate of revenue sharing by the Food Delivery App)
- Success rate of payments methods, based on regions (possible to find out with the dataset
provided)
- Card Expiry date might be relevant to calculate percentage of fraud happening because of
expired date (not enough information to find out)
27. PART C
In this section, I will be discussing the Wallet and Combine Features further. After which, I will try to
prioritise which feature is recommended to be developed further.
At this stage, I will move ahead with 3 features in total. They are:
Feature 1: Split
Feature 2: Wallet
Feature 3: Combine
Feature 1: Split (recapitulating)
When a user is ordering for a group, during payment, he (/she) can split the payment, by adding
friend’s Accounts. Result is, total bill splitted and charged from the individual accounts.
This way, their accumulative delivery charge is reduced, and user experience enhanced. It also re-
enforces the idea of “account” and encouraged to be used for collecting “Wallet” (to be introduced
later) Tokens and Credits.
Figure 30 : Wolt Split Feature
Feature 2: Wallet
Wallet helps users to store credit amount in their account at beginning of each month.
In many places, specially like Poland, consumers often don’t have enough money to be paid for an
immediate transaction. Moreover, Poland is a very cost-conscious country. Consumers often make
calculation of their monthly expenses [22] and follows it through. In such cases, it becomes a useful
to have a feature like Wolt Wallet.
Step 1: Wallet allows you to store credit (amount) at the beginning of the month.
Step 2: The money is transferred from your primary bank account.
Step 3: Every Payment you make through the app in that month, is paid from your Wallet.
28. Figure 31: Wallet Feature
Feature 3: Combine
Many times, in countries like Poland, consumers don’t have enough money to make an immediate
transaction. User might want to combine 2 cards or the option of paying with cash (“Cash on Delivery”)
along with a card. We should have a feature which allows the flexibility to pay combining two kinds of
payment methods.
Use Case 1: User is paying with 2 different cards.
Use Case 2: User is paying a part with his card and remaining through cash on Delivery.
Use Case 3: 2 Users ordered together. User 1 and User 2 is paying with their own cards. This is useful
when Wolt Spilt cannot be used because User 2 does not have a Wolt Account.
Use Case 4: 2 Users ordered together. User 1 is paying with his(/her) own cards and User 2 is paying
with cash. This is useful when Spilt cannot be used because User 2 does not have a Wolt Account.
The following are the options, usually consumer has at their disposal while making a payment in
Poland.
29. Different Online Payment Methods
Figure 32 : Different Payments Method in Poland [4]
We will be considering only the options with “Pay Now”. Ideally, we should have atleast one from
“Cash Based” and “Bank’s Intermediation”. We might consider the option of “Cash on Delivery” (COD)
along with payment through Bank Card/ Digital wallet (like Google Pay).
Figure 33 : Combine Feature
30. Feature Prioritisation
Features to be build are: Split, Wallet and Combine. We need to prioritise on which needs to be built
according to priorities.
The factors I have considered to understand the priorities are:
- Feature Size and Ease of Engineering: Complexity of a particular feature to be built
contributes to the difficulty quotient. With high ease, our priority tends higher as it needs
lesser days, number of engineers and manhours. During actual work within Wolt, I will be
discussing with the Product Owner or Tech Team Lead to understand the Engineering Effort.
- Revenue Impact: a feature will bring in particular amount of revenue. Higher the revenue it
brings (revenue impact), higher is our priority.
- Customer Impact: higher it enhances the customer experience and empathises with their
context, higher is our priority for it.
Factors Multiplied by are:
- Ease of Engineering (0.25): as it is important but for a well-funded team, we can put it at lower
priority. If Wolt was at the beginning of their journey, I might have considered factor 0.35.
During actual work within Wolt, I will be discussing with the Product Owner or Tech Team
Lead to understand the Engineering Effort.
- Revenue Impact (0.4): This is a critical factor for any business, infact, any business is in
business because of potentiality of revenue. If Wolt was at the beginning of their journey, I
might have considered factor 0.5.
- Customer Impact (0.35): We are in business because we are trying to solve pains for our
customers but considering other factors critical for developing a product, I have kept the
Customer Impact factor at 0.35.
Result: Wallet to be developed further.
Wallet is a critical feature to be developed for Payment Experience. Especially in countries like Poland,
where consumers often don’t have enough account balance to make an immediate purchase.
That is why, Digital Wallet Service providers, like PayU have created 30 days pay later service, which
is can be used by 42% of the consumers [13].
31. Work Package 4: Wallet
User Persona
Figure 34 : Ewa, User Persona
User Story Template
Figure 35 : Wolt Wallet User Story
32. Wallet Product Requirement Document (PRD)
Figure 36 : Product Requirement Document for Wallet
Wallet Development Flow
Figure 37 : Wallet Development Process Flow
33. Development Effort Cost
Considering, the team composition as:
Expertise Specialisation Number
Design UX Prototype + Research 1
Development
Front-End 1
Back-End 2
Data Analyst Market + User Feedback Research 1
Product Manager Product Vision + Backlog + Scrum 1
Total Team Size 6
We can assume that the development effort will take nearly 3 months to complete. The development
breakdown is:
Figure 38 : Development Timeline for Wolt Wallet [23]
Expertise Specialisation Number
Approximate Salary
(€)/ Month
Weekly Cost
(Salary) (€)
Effort
Weeks
Effort Cost (€)
Design
UX Prototype +
Research
1 4500 1125 3.5 3937.5
Development
Front-End 1 5000 1250 2 2500
Back-End 2 6500 3250 5 16250
Data Analyst 1 4500 1125 2 2250
Product
Manager
1 5000 1250
12 (60%
Capacity)
9000
Total Team Size 6 Total Development Cost 33937.5
The total cost for developing the Wolt Wallet feature is: € 33,937.5 (+ 5000 € Misc.)
34. Scrum Method for Managing Product Backlog
Figure 39 : Scrum Methodology for Development Process [24]
The development efforts should be followed through a Scrum Methodology. After features
prioritisation, Product Backlog will be decided during Sprint Planning.
Core Sprint Week will map onto the core development week which will be of 5 weeks (approximately).
Moreover, the Product Manager should work in close loop with the Scrum Master, in this case it will
be the Tech Lead.
35. References:
[1]: https://en.wikipedia.org/wiki/Wolt#Timeline
[2]: https://en.wikipedia.org/wiki/Wolt#Timeline
[3]: https://www.workinestonia.com/estonia-is-best-country-in-the-world-for-digital-life-new-
internations-2019-report-says/
[4]: Polasik, Michal & Fiszeder, Piotr. (2010). The acceptance of payment methods on the Polish e-
commerce market.
[5]: European Central Bank, ‘Fifth report on card fraud, September 2018.’
[6]: Oxley, J. E. and Yeung, B. (2001), “E-Commerce Readiness: Institutional Environment and
International Competitiveness”, Journal of International Business Studies, Vol. 32, No. 4, pp. 705-723.
[7]: 2019 J.P. Morgan Global Payments Trends Report – Poland Country Insights.
https://www.jpmorgan.com/merchant-services/insights/reports/poland.
[8]: countryeconomy.com
|https://countryeconomy.com/countries/compare/hungary/poland?sc=XE15.
[9]: CEIC Data: Hungary. https://www.ceicdata.com/en/indicator/hungary/annual-household-
expenditure-per-capita
[10]: CEIC Data: Poland. https://www.ceicdata.com/en/indicator/poland/annual-household-
expenditure-per-capita.
[11]: Poland – Why is it a good ground for discount? Foley Retail Consulting.
https://www.foleyretailconsulting.com/poland-good-ground-discount/
[12]: Ecommercewiki.org, November 2017. ‘Poland B2C Ecommerce Country Report 2017.
[13]: Payu.pl, January 2018. ‘What will 2018 bring for the e-commerce sector in Poland?’
[14]: Online payments trends in Poland by Tap2Pay. https://tap2pay.me/online-payments-trends-in-
poland/
[15]: Innovations in Identity, Trulioo Blog. https://www.trulioo.com/blog/payment-fraud-
management/
[16]: Payment Tokenization Explained. Square. https://squareup.com/us/en/townsquare/what-does-
tokenization-actually-mean
[17]: EU law will affect how you pay with Wolt starting Sep 14. Wolt Blog Post.
https://wolt.com/blog/hq/2019/09/10/eu-law-will-affect-how-you-pay-with-wolt-starting-sep-14/
[18]: 3D Secure Explained. Sage Pay. https://www.sagepay.co.uk/support/28/36/3d-secure-explained
[19]: What is 3D Secure? Securion Pay. https://securionpay.com/blog/3d-secure/
[20]: Calendar October 2018: https://www.calendar-365.co.uk/calendar/2018/October.html
[21]: User Interview 1: Conversation with a Polish Friend working at Polish Embassy in Mumbai, India.
[22]: User Interview 2: Conversation with a Polish friend working with Slush and living in Helsinki,
Finland.
[23]: Wolt Blog. How we design and engineer product features at Wolt: Pre-ordering food.
https://wolt.com/blog/hq/2017/02/07/design-engineer-product-features-wolt-pre-ordering-food/
[24]: Scrum Process. https://blog.4geeks.io/scrum-for-digital-product-development/