eselling on täällä! signaalit ja oppivat algoritmit myynnissä
TRANSCRIPT
eSelling on täällä!
Signaalit ja oppivat algoritmitmyynnissä ja asiakaspalvelussa
Prof. Petri Parvinen, Ph.D.
Literate learner
Generalize w/ care
Open mind
TOISTO TOIMII, KUNHAN SISÄLTÖ VAIHTELEE
ASIAKAS KÄSITTÄÄ MAKSIMISSAAN KOLME HYVÄÄ SYYTÄ OSTAA
JOS EI PANOSTA RIITTÄVÄSTI, EI TULE MYYNTIÄ VAAN VAIN AIKEITA JA TUNNETTUUTTA
Muutamia psykologisia ”totuuksia”
HYVÄ MYYNTITYÖ JA ASIAKASPALVELU ON VAIKUTTAVINTA, MUTTA EI TEHOKKAINTA MARKKINOINTIA
TYÖ KANNATTAA JAKAA IHMISTEN JA ROBOTTIEN KESKEN
Signaalien pitääkin tulla muilta asiakkailta
Jutun juuren täytyy tulla digistä
Pöyry & Parvinen, 2016
+ Comparison between Whatsapp
shares, personal Facebook
referrals, paid posts, organic
search share button
+ Whatsapp vs. Facebook msg =
no difference
+ Whatsapp/FB = 4,2x search
engine
+ Whatsapp/FB = 3,1x search
engine
+ 8x more likely to share when on
mobile!
Social referrals generate 4x referral propensity(Köster et al. 2017)
Schneider, 2016
Sensors, please
+ Predictive
+ Reactive
+ Automated
+ Analytical
+ Optimized measured
effect
US $0,98 / piece, Free Shipping,
minimum order 1 piece
CASE: PONSSE 360 VIDEO
CASE: VR TIMBER AND SERVICE TRADING
Digikoppia ottava myyntiälykäs
myymälä
Types of data available12
Owned data Open data
Freely available
and not
controlled by
anyone
Paid-for data
Open data
controlled by an
organization
Paid data
controlled by an
organization
Panel data,
commercial
databases
Not tracked or
not in
accessible form
Routinely
tracked or
collected
Selling organizations: What are the
current, emerging and future best
practices in selling data?
Buying organizations: What are the
current, emerging and future best
practices for buying data?
RESEARCH
FOCUS
Some examples of early owned-data based business models (Pöyry &
Parvinen, 2017)
• Signaling service – real time feed for timing operations
• Trend prediction and alerts, cf. social media analytics companies
• B2B data sharing economy cf. central associations/unions (“who is moving where”)
• Credit score business
• Intention-based marketing
• Purchase avatars (“I am currently buying X and Y”)
• Corporate data room for sale
• Plug into our logistics network
• Plug in with your after sales
• Preferred partner service based on availability information (“get before runs out”)
• Social circles information (“your kind of people are going X and doing Y”)
• Money-for-my-recommendations/network
• Selling store-specific information to vendors
• Reselling paid-for KIBS information (consulting reports, market research, etc.)
• Outsourced / managed service model based on data classification
12.10.2017
13
Continuous balancing of all continuous variables
Individualized pricing is the norm
+ Costs θ of the weighted randomized probability matching algorithm are 291.88, giving a profit of 88.2% of Φmax. Costs of batch estimation are 364.37, 85.3% of Φmax.
+ WRPM significantly outperforms batch, t = −5.338, p < 0.001.
+ Additionally, the costs of WPRM decrease over increasing market size M
batch estimationWRPM
Time for superaffiliates?
Social selling grows direct:
Could a customer avatar answer
customer questions on your
website live right after the
purchase?
Crowding, out-of-office work, two professions, globalization of saleswork, combining backoffice and front office, B2B e-commerce, salesifying customer service and service processes
