Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Razorfish & App Engine - A Modern Love Story

App Engine provides Razorfish with a platform to innovate at pace, be flexible throughout production, and deliver at scale. Hear our love story of how we've empowered teams to not only adapt to change but drive it too, using the Google Cloud Platform
by

Stuart Thorne

on 29 May 2014

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Razorfish & App Engine - A Modern Love Story

Razorfish & App Engine
A MODERN LOVE STORY
PROBLEM
OUTCOME
CORE TEAM
Keep it simple
Small core team (8 people), multidisciplinary, senior
SOLUTION
TIMELINE
SUMMARY
KEY COMPONENTS
App Engine
Argos
Analytics
facebook
Prediction
API
December &
Post launch
Launch - Oct 16
September
August
OUR VISION
High expectations
BUSINESS NEEDS
Huge challenge ahead
A first for Argos - Digital Gift Guide experience
THE SOLUTION
In 3 months
Fast, flexible, scalable solution - App Engine our clear choice
THE BRIEF
Cloud
Storage
Adobe
SiteCatalyst
"Reimagine the Argos Christmas Gift Guide for a digital environment in just 3 months"
Don't interrupt the existing Roadmap of Argos or their SI
Create a transactional Tablet E-Tail experience for the web
Event based deadline - Christmas 2013 - failure is not an option!
Ambitious and provocative Argos brand representation
Rich 2D/3D inspirational shopping experience
Handle thousands of business data changes hourly, daily
Connect multiple data sources; Argos, Systems Integrator, Adobe
Work collaboratively with Argos & their Systems Integrator
Iterative project process
Powered up with smart use of business and contextual user data
Stuart Thorne
Creative Technology Director, Razorfish
START WITH DATA
App setup, Authentication and Permissions - Day 1
API access and other core business data received
Defined data entities and relationships
Interface to Argos' APIs to retrieve Product and Store data
Completed Week 1
AUTOMATE
CRON Jobs
Expecting new Products throughout development
Frequent Product re-pricing and re-indexing
Critical data backed up to Cloud Storage - built-in!
Task Queues
Rate limited Transactional API calls, during Dev
Pre-cached 000's of Product query combinations
Tests, Exceptions & Notifications
Email alerts e.g. for API failure, job completion

Self-sustaining App
CREATE QUERIES
Create Datastore queries across entities and relationships
Quickly discovered new opportunities for filtering, e.g.
Filter Products in ALL Categories by Gender, Age Group & Price - doesn't exist yet on Argos.co.uk
Query indexes are automatically generated - nice!
Implemented sharded Memcache layer on top
Fast, flexible
EXPOSE SERVICES
Connected Queries with Request Handlers
Defined data interfaces with Frontend
Implemented security measures, such as CORS restrictions
Exposed RESTful Web Services to (any) Frontend
Connected functional test suites
SOA approach
Pre 2010, production was split like this...
Time
Resources
Confguring
Creating
Speed
Flexibility
Scale
What we needed was...
Sound familiar...?
Easy
Scale X & Y
Modular
Ready
Collaborative
Choice
What has App Engine actually done for us?
Changed how we work
Truly rapid prototypes - hail the 24 Hour App
Proactive approach, rather than reactive
DevOps & Service Ops
Helped drive what we do
Showing, rather than telling helps us win work
App Engine powers 90%+ of our concept work
Empowered new expertise and exploratory thinking
Helped us create best in class examples of work, for many clients
GIFT FINDER
The Gift Finder tool is an inspirational way to filter Products. Users choose from a series of images that best represent their loved one's interests. Images are tagged against Products using a unique scoring system.
Adapt data for new features
PREDICTIONS
Is it working as expected?
DATA SEEDING
Clever workaround?
Challenge
Solution
Easily apply new attributes to existing entities
Re-index Datastore & update Memcache on the fly
Test & learn with our scoring system, and fine tune
Challenge
Solution
Present recommendations as Gift Ideas throughout the App, starting with the Product Detail View
Rapidly change Prediction Models, re-generate data
V1 too precise; finds known relationships
Refine and improve iteratively, within hours and days
We needed to seed the App with real business data, to make Trends and Predictions more relevant. Use Argos.co.uk analytics? Yes!

BUT, we couldn't programmatically access Adobe SiteCatalyst in time...
Challenge
Solution
Adobe AEM has an analytics mailer tool
App Engine has a built-in Inbound Mail handler - great!
Created a Backend task to receive, unpack and apply the data to Prediction models - this took just 2 hours!
"PANIC STATION"
A unique feature, data driven
Challenge
Solution
Create a neat utility for last minute shoppers, making use of Argos SoS. Add recipient's age & interests, set a guideline price, then view a selection of Products guaranteed to be in stock.
Argos SoS Home deliveries until late on Christmas Eve!
Connected real time Product availability
Targeted & relevant results - per User filter, per Store
FEEDBACK
HOW DID IT DO?
Busy: 1M visits, 5M page views at peak pre-Christmas week
Engaging: Average dwell time 3:05 minutes
Sticky: Repeat visitors dwell 25% longer
Valuable: 'Panic Station' increased our conversion 600%
The finished product
App Engine has grown as we have, over the past 4 years
TRENDS
Combined business data, analytics and User context (e.g. browse history, location) to surface trending Products
Cyclical process: User browsing behaviour analysis is used to feed Products back into Trend Boards
Created a Custom Admin Tool for curating Trends, e.g. for key merchandising opportunities
Smart use of data
Thank you!
Stuart Thorne
@razorfishlondon
Argos Digital Gift Guide
A Technical Case Study
Argos Digital Gift Guide
I want it now!
Time to market ever decreasing
We expect change, we drive change
No longer optional, it's required
I make mistakes ...
... then fix them
Quality is paramount
Remove effort, risk and worry
Capture and play back Consumer shopping trends in real time
A Powerful Family
Trend Board
Categories
Product Detail view
Basket
The role
Technology
Creative Thinking
Awesome Consumer Experiences
What is success going to look like?
Reliable: Low latency & zero downtime in App Engine application
Positive buzz in industry press, blogs and social...
Custom Recommendations Engine, inspiring gift selection
Why App Engine?
Full transcript