Building Data Driven Products using Snowplow Analytics

Digdeep | Analytics

Building data driven products using Snowplow Analytics

December, 2015

Today, your product/ customer matrix is overwhelmed with data being made available by all the connected devices out there. 

As a data scientist, you are tasked with defining where the growth opportunities are. For the company, the opportunities are real, and they are big. Data is the new currency. 

In this talk, I will walk through a use case in monetising your Snowplow data. The focus of this talk is about product development, in building transactional relationships between companies where both sides benefit from new products and services using Snowplow Analytics.

All your Digital analytics data, in the one place.

Customers |

Collect
Enrich
Access

The Growth Opportunity Matrix

Why Snowplow Analytics

A real world example

Agenda

The Growth Opportunity Matrix

Why Snowplow Analytics

A real world example

A Step-by-Step Guide for CMOs and CIOs

As Forrester says, “Customer obsession is easy to talk about but hard to do. It requires remaking your company, systematically, to re-orient each element toward improved customer experience.”

A blueprint for updating product strategy in today’s Age of the Customer, based on four key strategic imperatives:

Transform the customer experience with a systematic, measurable approach.
Accelerate your digital business future to deliver greater agility and customer value.
Embrace the mobile mind shift to serve customers in their moments of need.
Turn big data into business insights to continuously improve your efforts.

The growth opportunity matrix

5. Pay us lots of money and we will position you in the top right of our magic quadrant

The growth opportunity matrix

Product Development

Customer Penetration

Market Development

Diversification (Venturing)

A tried and tested blueprint for updating product strategy in today’s Age of the Customer, based on four key strategic imperatives:

Objective: Distinguish among 4 generic growth strategies by categorizing them along 2 intersecting dimensions based on whether they are “established” or “new” products and markets.

Each of the quadrants has implicit tactical implications, your data strategy should define when/ which quadrant you are in and the what/ how value you are offering.

The growth opportunity matrix

Customer Penetration - OLD / OLD

Increase profitable sales of existing products to customers already being pursued (but not necessarily buying at the current moment).

TGT Market (Penetration)	Increase number of user	Diffusion/ Adoption Models
Current Users (Usage Rate)	Increase Usage	News uses/ occasions
Consumption (Share)	Increase Share of customer purchases	Higher customer loyalty, Higher Margins
Sales	Increase customer lifetime value	BTYD

Related Frameworks

The growth opportunity matrix

Market development - OLD / new

Increase profitable sales of existing products to groups of customers who are not currently being pursued.

Geo Expansion	Customer Segmentation	Product Lifecycle

Related Frameworks

The growth opportunity matrix

product development - new / old

Sell a wider variety of products to current (existing) customers or upgrade products currently in use.

	OLD	NEW
NEW	REACT Follower strategy	INNOVATION (unrecognised needs, New Tech)
OLD	REFINE (incremental improvement)	REPOSITION/ "PIVOT" (new uses, new contexts)

Market

Biz

The growth opportunity matrix

diversification - new / old

Risky business. This is hard, let's leave it at that.

</too hard>

The growth opportunity matrix

Does my data support a particular product/ market strategy?, and can my data be the product strategy?

The Growth Opportunity Matrix

Why Snowplow Analytics?

Demo

Narbeh Yousefian

Focus on Product

Stay committed to your core product and leave the analytics infrastructure and maintenance to the experts. Snowplow make engineering analytics a non-issue.

Narbeh Yousefian

Tailored Insights

Accelerating decision lead times with automation, answer the business questions that actually matter with event detection.

Narbeh Yousefian

Capture Everything

All your data sits in the one place. Embed analytics anywhere and collect every interaction in batch/ real time using an event driven data model.

Narbeh Yousefian

Data Products

Profile, classify, score and expose your customer segments back into the application with custom API's. Target opportunities and send data with programmatic solutions and integrations.

An open framework,
flexible and extensible

Why snowplow analytics?

Narbeh Yousefian

Capture Everything

case study: shifting from a stagnating business model to a growth strategy in the property vertical by using snowplow analytics to drive a “new” product in an existing market.

product development - new / old

Why snowplow analytics?

select a.domain_userid, a.event_id, a.geo_city, a.geo_latitude, a.geo_longitude,
b.address_id, b.latitude, b.longitude, b.postcode, b.state_code, b.suburb, b.on_market_type, 
2 * 6373 * asin(sqrt((sin(radians((latitude::float - geo_latitude) / 2))) ^ 2 + 
	cos(radians(geo_latitude)) * cos(radians(latitude::float)) * 
        (sin(radians((longitude::float - geo_longitude) / 2))) ^ 2)) as distance
from (
select distinct domain_userid, event_id, geo_city, geo_latitude, geo_longitude
from a.b
where event = 'page_view' AND geo_city IS NOT NULL AND geo_latitude != -27 AND geo_latitude IS NOT NULL
AND ( br_type = 'Browser' OR br_type = 'Browser (mobile)' )
AND user_ipaddress != 'x.x.x.x'
and exists (
select device_id
from (
select distinct device_id,
count(distinct(domain_userid)) as count
from a.b
group by 1
order by 2 desc
)
where count < 20
)
AND exists (
select event_id
from (
select distinct event_id,
count(event_id) as count
from a.b
group by 1
order by 2 desc
)
where count < 10
)
) a LEFT JOIN (
select distinct root_id, address_id, latitude, longitude, postcode, state_code, suburb, on_market_type
from a.c
WHERE longitude != 'x' AND longitude is not null AND longitude != 'undefined'
) b ON a.event_id = b.root_id

Your data, you own it. Raw access to SQL, scaleable and extensible columnar database.

Narbeh Yousefian

Tailored Insights

Quantile	0%	25%	50%	75%	100%
Distance	6km	12km	41km	672km	1,889km
Users	12,406	12,550	12,575	6,109	6,360

Sample == 50K users

Why snowplow analytics?

Narbeh Yousefian

Tailored Insights

Blending geo centric data streams with 3rd party data sources.

Why snowplow analytics?

Narbeh Yousefian

Data Products

Made with

OPEN street maps + leaflet.js + mapbox

Blending geo centric data streams with 3rd party data sources.

Why snowplow analytics?

Narbeh Yousefian

Data Products

data = [{
              'data': {
                'user': {
                  'UID': ..., 'DID': ..., 'member_id': ..., 'member_type': ...
                },    
                'geoLocation': {
                  "city": ...,
                  "continent": ...,
                  "latitude": ...,
                  "longitude": ...,
                  "country": ...,
                  "deviceTimezone": ...,
                  "region": ...
                },
                'Affinity': {
                  "type": ...,
                  ...
                },
                'Distance': {
                  "AVG_distance": ... , "MIN_distance": ... , "MAX_distance": ... , 
                  "STD_distance": ... , "MDN_distance": ... 
                },
                'Price': {
                  'Buy': {
                    "AVG_price": ..., "MIN_price": ..., "MAX_price": ..., "STD_price": ..., "MDN_price": ...
                    },
                  'Rent': {
                    ...
                    },
                  'Share': {
                    ...
                    }
                  }}}]

Recommendations, content targeting and experiments - power all your services with an open architecture to build and expose data products through RESTful API's.

Why snowplow analytics?

Narbeh Yousefian

Data Products

The Growth Opportunity Matrix

Why Snowplow Analytics?

Demo

Digdeep | Analytics