A talk I hosted at the Business analytics Insight seminar in Utrecht (NL) and Diegem (BE). The focus was on various options to accelerate time-to-market for innovative business analytics projects.
2. Progress in analytics comes as an
evolution, rather than a revolution
High value business cases take management analytics
beyond data summarization.
Trend to shift expenses towards OPEX.
Challenges in mid-term planning due to
rapidly evolving technology landscape
3. Often in a climate of
risk aversion and
skill gaps
4. On the API
economy:
Automated data sharing is
a central feature of
multiplayer innovation
Source: economist.com
Diversification in demand
from analytics teams lead to
global ecosystem of
standardized,
off-the shelf services from
both generalist and specialist solution
providers
Service offerings package typical
specialist workload in
modular and
interchangeable building
blocks
5. API’s offer on-demand access to
3. Algorithm
markets
Off-the-shelf and purpose-fit
solutions for highly complex
challenges
1. Data storage
and processing
Major players commoditize
upfront data analytics
innovations
2. Specialist
data feeds
Frequently updated and hard
to obtain data is available via
subscription
6. 1. Data storage & processing
➔ On-demand solutions for complex data challenges
Thanks to economies of scale, cloud solution providers manage to dramatically lower the entrance
barrier to complex data challenges.
➔ Web-scale companies lead innovations
Globalized, ‘open-sourced’ innovation models help making web-scale companies profitable.
➔ Notable use-cases
Realtime and complex analytics on large datasets
Proof-of-concepts and time-to-value optimization
Cost rationalization by automation
7. 2. Specialist data feeds
➔ Hard-to-obtain data is commoditized and available via subscription
In a bottom-up approach to generating business value, detailed data is of high value although
expensive to gather and manage internally. Specialist brokers focus on commoditization.
➔ Data brokers typically sell fast or aggregated data, with exceptions
Business model for many Social startups.
Cloud-first software vendors allow data access via exact same API principles.
➔ Notable use cases
Integration of weather data into planning processes
Use of social data in sales analytics
Access internal data from
8. 3. Algorithm markets
➔ Complex algorithms are packaged in an on-demand offering
Skill gap often leads to delayed adoption of academic R&D in business environments
On-demand commercial offerings apply insights on proprietary company data
➔ Algorithm markets gather building blocks from independent firms
Many academic spin-offs have API’s available yet don’t focus on marketing.
Trusted web-scale vendors either market or implement fundamental research in a commercial offering
➔ Notable use-cases
Automating repetitive tasks on unstructured data
text analytics, image recognition, market forecasting
9. Build an end-to-end analytics pipeline using
API’sData Storage
Use Amazon API’s to
ingest real-time data, at
scale.
Data processing
Plug managed Apache
Spark into AWS to
transform data at scale.
Licensed analytics
Connect to 3rd party API
vendors/consultants to tackle
specific boxed tasks.
In-house ML/analytics
Use on-demand capacity to
fuel model training on
properietary data.
Idea Go-live
Insights storage
Purpose-fit analytical
databases grow along
with the project.
Cloud-based
consumption
Consume insights either by
building your own API or by
using cloud reporting
platforms.
10. On the API
economy:
Automated data sharing is
a central feature of
multiplayer innovation
Source: economist.com
In an API ecosystem, analytics teams
are enabled to tackle more
complex challenges in
less time by using
commoditized building
blocks
The key to success is
integrating existing
analytics efforts with
innovative API’s rationalizes time-to-
value and reduces risk