This is an example case study showing what big data can mean for a small website that generates just 5000 visits a day.
It all depends on what we want do get from our assets like website traffic. If we only measure the number of people who visited our site, then we do not need to worry about “big data”. We just have to count total visits (5000 a day, 150 000 monthly).
But by using just the simple measure we know nothing about our visitors / customers. So, it pretty useless.
On the following slides we present what a website owner can gain from advanced website analytics and why big data technologies are recommended.
2. This is an example case study showing what big data can mean for
a small website that generates just 5000 visits a day.
It all depends on what we want do get from our assets like website
traffic.
If we only measure the number of people who visited our site,
then we do not need to worry about “big data”. We just have to
count total visits (5000 a day, 150 000 monthly).
But by using just the simple measure we know nothing about our
visitors / customers. So, it is pretty useless.
On the following slides we present what a website owner can gain
from advanced website analytics and why big data technologies
are recommended.
Website analytics – Introduction
3. Website visit – Example flow
Website visit
Landing
page
Search
page
Product
page
Product
page
Source page
(referrer)
Customer
leaves
our website
Each of these 5 events is:
collected, enriched with additional features/dimensions
and stored for further analysis and processing.
4. Website visit
Event type Number of
data points
Total number
of data
points (sum)
Website visit event (1st page view) 1 1
Average number of page views during one visit 4 4
Where the customer came from (“referrer”) 1 5
Raw data from 1 event
(IP, URL, cookie, user-agent, timestamp)
6 30
Enriched dimensions* from 1 event
(product, behavior, device, location, etc.)
50 280
A customer visits a website…
280 information from 1 website visit gathered.
* See Appendix for the list of possible website data enrichments.
6. Website visit – data volumes
Time span Number
of data points
1 event 280
1 day (5000 events x 280 data points) 1,400,000
1 month 42,000,000
1 year 504,000,000
Let’s assume, we have 5000 website visits daily
42M monthly, and over 500M annually data points collected.
8. Let’s assume we want to find users who:
Were interested in buying healthcare insurance
Use Apple product
Live in cities with population over 1M people
Are woman
Came from our display campaign.
So, we have a combination of 5 (k) dimensions from 50 (n).
Using the combination formula: we will have…
Website visits – analytics
11. So, we will analyze:
millions of data points daily, and hundreds of millions annually
in thousands of possible ways
on data that is streamed in real-time and that may change its
structure in time
That is why these standard methods may not work:
Non horizontally scalable systems (like legacy relational databases)
Data aggregates (it is not possible to implement all combinations of
data aggregations)
Relational databases with fixed, pre-defined data schema
Website analytics – summary
12. And one more thing…
That was a simple task
of analyzing our website traffic only.
13. The real value is in combining the other data, like:
Product usage
Mobile applications activity
Data from devices (IoT, like: beacons, car plugins etc.)
Marketing campaigns
Other customer touch points like: phone calls, emails etc.
This adds tens if not hundreds of other dimensions and multiplies
number of events.
Try your math skills and do the simulation by yourself :)
The real value: adding more data sources
14. User/customer segmentation
Product recommendations: up-selling, cross-selling, next best offer,
next best action
Content recommendations
Ad recommendations
Conversion optimization
Retention optimization
Dynamic prices
…
How we can use this data – examples
15. Share your thoughts, challenges
or case studies with us.
We help high-growth companies
to make real-time decisions based on big data
by providing scalable, flexible and real-time
data collection, storage and analytics.
Or drop us a line: hello@deep.bi
SUBMIT»
17. Web & Ecommerce Data Enrichment
Location & Internet connection
LOCATION
• Country
• Region
• City
• ZIP Code
• Population
• Latitude & Longitude
• Time zone
• IDD prefix to call the city from
another country
• Phone area code
• Mobile Country Code (MCC)
• Mobile Network Code (MNC)
• Elevation
• Weather at the moment of event
INTERNET CONNECTION
• ISP name or Organization name
• Organization type:
• Commercial
• Organization
• Government
• Military
• University/College/School
• Library
• Content Delivery Network
• Fixed Line ISP
• Mobile ISP
• Data Center/Web Hosting/Transit
• Search Engine Spider
• Reserved
• Mobile brand
• Net speed
18. Web & Ecommerce Data Enrichment
Device, OS, Website content & Products
DEVICE & OS
• Device Type
• Device Brand
• Device Model
• Device Operating System
• Operating System Producer
• Browser
• Browser Producer
WEBSITE & KEYWORDS
• Domain
• Website category
• Keywords (semantic concepts)
• Unified product categories
• Product brands
PRODUCTS
• Product name
• Unified product category
• Store product category
• Keywords (semantic concepts)
• Product brand
• Product regular price
• Product sale price
• Price range
• Product features
19. Web & Ecommerce Data Enrichment
Demographics, Time & Behavior
DEMOGRAPHICS
• Gender
• Age
• Occupation
• Martial status
• Children
• Earnings
• Price sensitivity
TIME
• Event Timestamp
• Day part (morning, noon, etc.)
• Day of week
• Day of month
• Month
BEHAVOR TYPE
• Page view
• Page section view
• Page attention time spent
• Page section attention time spent
• Ad view
• Ad view time
• Ad click
• Ad engagement (hover, etc.)
• Search
• Product view
• Adding to cart
• Purchase
20. Share your thoughts, challenges
or case studies with us.
We help high-growth companies
to make real-time decisions based on big data
by providing scalable, flexible and real-time
data collection, storage and analytics.
Or drop us a line: hello@deep.bi
SUBMIT»