digieco Open Seminar - ict Competitive Strategy
ETRI Researcher Lee Seong-jun
- Telco era
externality (externality/external effect)-based shifts in value, utility
Examples: subway, mobile-phone subsidies, KakaoTalk/Facebook sign-ups (sign-ups create more sign-ups),
Economies of scale and natural monopoly, One policy, universal service
rate of return regulation(Bell lab), stand-alone method regulation
- PC era
Externality of indirect (virtual) networks (influence via systems, OS)
Example— compatibility) IBM and MAC ripple effects by user count(h/w+s/w),
Software sells hardware. The floating-point dispute (a problem Wozniak couldn't solve — collaboration with Bill Gates to fix a technical issue in spreadsheets gave the opening for information leakage), locked in to a poor service (inferior good), critical mass point, monopoly — we must watch the multiple equilibria before and after the critical mass point (demand and supply match), Dvorak is far better than QWERTY but — efficiency aside — since one ecosystem dominates, people don't adopt it.
Even attempts to shift from QWERTY are rejected, path dependence!!: train rails, the width of carriage tracks, Roman-era fitted-to-all standards — the impact one standardization has on an entire business sector (=> the power of my HTML5!)
Expansion on a subscriber base, compatibility dominance, tolerance for piracy, manipulation of expectation!
comparison spreadsheet ratings
A result that, no matter how much you wanted to side with MS, there was no choice.
- post-PC era
Platform — a place for common use, a structure to access work or an event, L-shape: expanding the subscriber base via free, etc. (-> sell subscriber info, open APIs, expand horizontally via platform), fundamental condition of a platform: generalization!, middleware (like browsers — not controlling the whole computer like an OS, but universally used), BIOS (IBM-compatible BIOS manufacturing boom — generalized hardware), *order of generalization: hardware -> BIOS -> basic software (OS) -> middleware -> application
- cloud era
speed:distance = time:space, Gilder's law
Cloud is consumed on the same concept as electricity.
Google: thin client = Apple: expanding device functions(app)
Scalability, cost innovation, real-time, mobility, increased number of devices (difference from the old mainframe era).
big data & paradigm shift
N1=1all full enumeration(not sampling,)
Sometimes quantity matters more than quality (Google Translate / previously, elaborate systems / reduced to probability problems), wiki style, lowered data-collection criteria.
causation -> correlation(prediction, new discoveries), the end of theory(the death of experts), pre processing -> post processing, we can find data paths that predict results without knowing the cause(important but risky... the future homework of big data / spot the keyword first, then craft a story saying it's selling the phenomenon to persuade others. People want packaging)
API - digital Glue or a contract
Microsoft used it first.
Bundle APIs sdk release -> product-scope expansion -> customer/manufacturer expansion
External network effects! (Facebook connecting to 1+ million sites, etc.)
Strengthening the base for KakaoTalk-capable subscribers / Twitter universe.
Algorithms — how big data makes money.
Turning very complex problems into concrete calculation procedures whose order is clearly defined, simplifying them into very simple structures (reconstructing complex problems into yes/no, example = flowchart), the essence of an algorithm is prediction, but you need lots of data. Big data business models, extended generalized second price auction, Adsense, recommendation systems(Netflix — 60% of revenue), auto complete
- Future outlook
chromcast(smart TV), proj.ara(handset) -> generalization!
The more internet users, the more they themselves are the revenue model. But manufacturers sit on 1-million-won-plus products.
Black-boxing core tech + opening the platform (Genichiro Senoo).
Android fork (customization) response.
-> Google locks the ecosystem into Google apps, not the Android OS!
-> A strategy to platformize APIs rather than the OS
More devices per person, wearable devices,
Multi-OS, m2m, iot era's
'main secretary'
Among Apple, Google, and Samsung — all of whom own OS-level service platforms — who will play the hub is the important emerging issue.
the making of a fly (a blind spot of algorithms — while the customer sits still, sellers auto-bid among themselves, pushing a book to a million dollars), flash crash(downside of algorithmic trading systems)
In the past, MS was able to generalize-monopolize by setting standards with Intel.
Does Google have a strategic means to prevent future Google apps from being generalized?
Samsung is combining content in cooperation with operators in each country.
Smartphone — the lead of touch UI — + wearable devices.
-> Limits in input/output methods (-> next-gen UI: voice, motion, eye gaze, brainwave, hologram output...)
: There's a gap between new technology and human usage; at that intersection lies business opportunity...
(The gap between MP3, Soribada, and Napster was shrinking.)
datafication, turning everything into data ( ex - beddit )
Netflix
House of Cards content (Doesn't stream every movie — The Dark Knight Rises isn't there. Uses internal data to predict per-movie profitability up front.)
API specialized companies
In the big-data mining process, what is the human role, and who solves the problems algorithms generate, and who is responsible for the issues they cause?
When the wind blows, the candle dies, but the bonfire grows.
The way to survive in the coming ICT ecosystem competition is to become antifragile.
When uncertain issues hit a well-tamed system that is standardized, averaged, and stripped of uncertainty, becoming antifragile is the way to survive. Otherwise it can be broken easily by external shocks.
- Questions
Shouldn't we also think about content beyond the platform?
