2. • Mobile Developer - pretty much every platform!
• Lecturer in Distributed and Mobile Computing at ITT Dublin
• UCL - Tech Lead at Dept of Innovation and Enterprise
• Tech Trainer
• Mentor at
Background
3. Accepted 500 companies to date
AcquiredJobs created Official twitter partner
@ucladvances
MVP
Committed
Team
4. UX LAB DEVICE LAB APP LAB
UCL Digital Business Labs
@ucladvances
5. Gather detailed use analytics to learn and
iterate with the help of our UX Consultants
and connect directly with users hoping to
become App Ambassadors.
Beta test in App Lab, UCL's app
store exclusive for it’s 45,000+
staff & students
APP Lab
@ucladvances
6. Project TRIANGLE
• Project TRIANGLE is a future internet testbed (FIRE+) which
prepares the ground for 5G certification.
• In the project we create a pilot platform for ‘pre-standards’ ‘pre-
normative certification’ of 5G devices and 5G applications.
• During the project we will run a number of experiments at the device
and app level and interconnect other testbeds and technology to the
platform
• Budget ~3M, EU H2020 funded
• Open Calls for SMEs 500k euro funding
• Website - http://www.triangle-project.eu
7. Overview
• Startup Company Growth Stages and needs
• Metrics and Measurement for Investment
• SME Considerations for Architecture
• Case Studies
• Workshop
8. disruptive business models
• Models
• Long tail for retail
• Sharing Economy
• Marketplaces
• On Demand Services
• AI and Machine Learning is causing further disruption
9. disruptive business models
• Value creation and destruction
• Traditional approach to make your own products
obsolete
11. On Demand
• Uber
• Deliveroo
• People as a service
• Task Rabbit / Ginn / Washio
12. Marketplaces
• Displacement of traditional intermediaries or agents
with market place
• Examples
• Ebay Amazon
• Hassle
• Appear here
13. Stages
• Idea generation
• Business model canvas
• Interviews
• Mockups
• Feedback and user testing
• Identity pain points and solutions and routes to achieve
• Product solution fit
• Product market fit
• Scale
17. Innovation Accounting
• Measuring and validating assumptions using data
• Enables key decisions to be made using proof from
data
• Build - Measure - Learn cycle
• Using sprints to validate assumptions
• Good Analytics are key to implementation
18. MVP
• smallest solution that delivers (and ideally captures)
customer value
• identify assumptions to validate
• Concentrate on critical path
• Techniques
• golden path
• wizard of Oz
20. Metrics
• identifying target users
• cost of customer acquisition
• average revenue per user
• monthly recurring Revenue
• average session length
• bounce rate
• monthly active users
21.
22. Important Investment
Metrics
• What is the funnel -
• inbound leads, conversion rates, drop off rate, growth rate,
• retention (MAU), referral, churn
• size of addressable market, burn rate
• SaaS
• cost of user acquisition, lifetime value of customer, APRU
• MOS NPS
23. SaaS Metrics
• ARPA (Average Revenue per Account per month)
• Net MRR (Monthly Recurring Revenue) (including MRR expansion)
• LTV
• CAC
• LTV: CAC ratio (should be >3X)
• Months to recover CAC. (should be <= 6 months)
• Customer Engagement Score
see http://www.forentrepreneurs.com/saas-metrics-2/
24. Architectural Decisions
• Tradeoff between speed, suitability, performance
• Common approach is to build a rough demo that
won’t scale, then rewrite
• some accelerators (500 startups) will not look at a
startup that is not using cloud to scale quickly
• cloud cost to demand forecasting needed for
metrics and pricing cost of running and scaling cost
to support next N users
25. When to Build
• build vs buy / use / adapt
• we have smart people, we can do that (better)
• you might, but should you?
26. Build vs use
• concentrate on USP not the surrounding building blocks
• technology blocks are a commodity, compute is almost free
• contrast to dot com boom when more needed to be built
• speed to traction is more important than the underlying tech
sometimes
• business model can trump better technology
• doing it right vs best
27. What to Build
• Early stage companies should focus on the core
differentiator
• and validating assumptions
• Use existing off the shelf components where possible
• Sometimes you don't need to build the entire
solution, just the key part
• Outsource or use existing resources
28. Secret Sauce
• value of defensible IP
• IoT China, need to be first and best,
• get a community GoPro is not the only camera)
• proprietary algorithm - better than what is now in public
domain Google/Microsoft/Facebook/algorithmia?
• more efficient approach vs throw money and hardware at
the problem
• building a data lake important for investors right now
29. PaaS
• Parse - for many years one of the most popular options for App Developers
• AWS & Lambda
• Firebase
• CouchBaseAWS Elastic BeanStalk
• Google App Engine
• Heroku
• IBM Cloud Platform
• MS Azure App Services
• Oracle Cloud Platform
• Red Hat OpenShift
• Many many More..
30. Mobile Development
• HTML5 vs native
• Lots of tradeoffs
• Cross Platform tools
• Hybrid (PhoneGap)
• single codebase (ish) - multiple platforms
• React Native, Appcelerator, Xamarin
31. Data Data Data
• Data is the new currency
• many businesses are pure data plays
• need to understand the value of the data and how to extract it
• setting the correct architecture for mining data
• consider storing the features of the data in a format where they
can be trained
• pipelines for ML see Spark or AWS Lambda as examples
• video streaming realtime analysis from Facebook
32. new arch models
• Client/Server split is ever-changing
• Source of new innovation in the future
• Fog and Edge Side Computing
• Growth of IoT and small devices
• Differential Privacy and on device processing
• Network Apps
33. death of apps?
• Lots of talk of the decline of apps
• user engagement dropping
• 70% of user time is in some type of messaging
platform
• AI platforms are gaining in popularity
• Alexia, Cortano, Allo, Facebook
• Rise of the bots
34. Bot platforms
• NLP + context understanding are crucial
• Roll your own
• off the shelf decision tree
• Examples
• weave.ai
• blue mix + Watson
• Viv
35. Approaches
• Object Oriented Programming not only option
• Data Driven architectures, popular in the gaming community
• Data-Oriented Design
• http://www.dataorienteddesign.com/dodmain/
• http://gamesfromwithin.com/data-oriented-design
• Procedural - for machine learning C/Python
• Functional - Haskel/R/Lamda
36. Processing Data Streams
• System Architecture is adapting to new data requirements
• the proliferation of low cost Internet connected devices
generating lots of data
• Different data types from Social Media Analysis, Fraud
Detection, to Video Processing in Realtime
• Stream Processing and Streaming Analytics
• Machine Learning can make decisions on live data and
continue to learn
• Processing of large data sets in non-realtime
37. Popular Trends
• Trend towards open source and PaaS, BaaS and platforms
• Architectures
• Micro services
• MOM and Queue based architectures
• Stream Based - Spark
• Deployment
• AWS Lamda
• Open stack
• Docker
• Ruby Django Heroku
• Continuous Integration
• Travis / Jenkins
• Test Automation - Selenium / Appium / QuaMotion
49. value proposition
What is the most important thing to show/build to demonstrate your value proposition
peel back the layers of the Onion
What do you need to do to unlock the channel
critical path
MVP golden path and wizard of Oz
50.
51. Marketplace Issues
• Seeding marketplaces is a big issue
• demand side vs supply side
• Revenue models adapt to the value of data vs transitional or
advertising
• More data means more pricing predictions
• Importance of volume is huge to avoid competition as is the ability
to subsidise one side to allow growth
• PM fit functions, generates value for both sides of the
marketplace, business model scaleable and growth driver is right
52. Marketplace Issues
• 1) cold start
• 2) liquidity 100 applications and 10 things get done a
month (retention or referral rates, transaction rates)
• 3) leakage - go off platform direct and cut you out
• 4) revenue model - growth driver (1 what is best paid
acquisition - SM Facebook ads etc do this get that 2
retention strategy small loans to keep them, good
account management do same next month 3
53.
54. Product Testing
• Alpha and beta testing
• closed vs open betas
• growth hacking
• SEO SAO
• Using Social to promote growth and engagement
Notes de l'éditeur
criteria
- The programme that you are here to take part in is for earlier stage companies and focuses on validating and testing your product.
- Over the coming months you will receive access User Experience consultants, a device lab & App Lab
App Lab is a private app store that’s exclusively for UCL’s 45000 staff and students to access the latest apps… Those are the apps that you guys are building.
Normally when you test apps, Friends family alpha Chances are, there aren’t many of them and they’ll be biased. Your mum will tell you it’s brilliant whether it is or not!
Then through services like Google Play Beta Test Flight, you have to find or buy your own testers. App Lab gives you the chance to do large scale, un biased, beta testing. It brings you detailed analytics of how real users are using your app. you can then work with our UX team who can help you prioritise which changes will have the biggest impact. You can then iterate really quickly, we’re not apple! and then see the impact these changes make. Even if you already have an app in the public app stores you can use app lab to try new things without effecting your current use base. Private APIs.
It’s available on iOS - and it’s a pretty rare thing to have an apple app store not run by apple!! And we’ve JUST LAUNCHED an android version too.