Many large operators have expressed a desire to undertake disruptive change, and we have often proposed an agenda for such change. What typically happens is that, after several rounds of engagement, we observe that there is little mainstream organisational appetite to engage in disruption. Why so?
The main reason is a perception gap between the current state of the art (which any leading operator delivers) and our understanding of the state of the possible (which most operators are very far from). This gap exaggerates the risks of engaging in disruption, and underestimates the potential rewards.
Another reason is that our industry as a whole implicitly believes that network service quality is a matter of detecting and rectifying ‘faults’. This framing inhibits the consideration of the alternative paradigm of networks as resource trading spaces. As a result, the significant ‘quality arbitrage’ that exists in all IP networks is not visible.
Operators face the risk that others will exploit the arbitrage opportunity, to their serious commercial disadvantage. This has happened before, e.g. with TDM and the rise of ISPs, and is happening now with SD-WAN. We propose that a larger multinational operators need to proactively initiate the disruption via a new business unit.
2. 2
Network quality arbitrage demands a
commercial and technical strategic response
Many large operators have expressed a desire to undertake disruptive change, and we have often proposed
an agenda for such change. What typically happens is that, after several rounds of engagement, we observe
that there is little mainstream organisational appetite to engage in disruption. Why so?
The main reason is a perception gap between the current state of the art (which any leading operator
delivers) and our understanding of the state of the possible (which most operators are very far from). This
gap exaggerates the risks of engaging in disruption, and underestimates the potential rewards.
Another reason is that our industry as a whole implicitly believes that network service quality is a matter of
detecting and rectifying ‘faults’. This framing inhibits the consideration of the alternative paradigm of
networks as resource trading spaces. As a result, the significant ‘quality arbitrage’ that exists in all IP
networks is not visible.
Operators face the risk that others will exploit the arbitrage opportunity, to their serious commercial
disadvantage. This has happened before, e.g. with TDM and the rise of ISPs, and is happening now with SD-
WAN. We propose that a larger multinational operators need to proactively initiate the disruption via a
new business unit.
2
4. 4The ‘quality arbitrage’ business opportunity
Most telecommunications providers currently deliver data services that are much better than
required most of the time. This is in an attempt to avoid episodes of poor quality that cause
customers to complain and churn. Despite this commitment to over-delivery, poor quality still
occasionally occurs. These occurrences are treated as ‘faults’.
Maintaining this over-delivery is very costly, and thus creates a substantial arbitrage
opportunity. Whoever can ‘trade’ resources to better manage supply and demand (over the
appropriate timescales) can exploit this opportunity. This need not be the network operator!
It could be a customer or a rival.
We have measured the nature and scale of this arbitrage in other operators’ networks. We
have also constructed new services to exploit it. Based on this experience, we believe that
every major operator has the opportunity to create more performance-segmented products
(in terms of quality and reliability). The way to achieve this is to get better control over the
QoE risk.
5. 5
Appropriately managing the network resource trades enables both of these to be done at the
same time, i.e. to execute the ‘quality arbitrage’. This arbitrage offers a rare opportunity to
create a highly disruptive business model that extracts much more value from the
underlying data transmission assets.
Demand for QoE and willingness
to pay are both highly variable
In other industries, customers are prepared to trade cost for
uncertainty, and the products reflect the variability in
willingness to pay for QoE. For instance, airlines charge different
prices for standby, non-changeable and fully flexible tickets.
In telecoms, some customers are willing to take on more QoE
uncertainty for a reduced price; their traffic can be used to run the
network ‘hotter’. Conversely, by reducing QoE uncertainty, you
can increase the value of the service to other customers, and
charge a premium.
6. 6
Exploiting the arbitrage leads to a
demand-led ‘digital telco’ business model
To exploit the arbitrage you need to make more ‘good’ resource trades than ‘bad’ ones than is
currently the case. This means you must consider a network as a resource trading space. In
this paradigm, a network is more like a commodity futures trading platform than a ‘pipe’.
In telecoms, the commodity’s supply is ephemeral, whereas demand can be deferred (from
microseconds to months). You make trades by time-shifting (and space-shifting) the demand.
This in turn requires you to predict the impact of a trade on QoE, and hence willingness to
pay. Predicting the effect of a trade requires you to appropriately characterise the demand
for QoE and hence performance (at least in broad terms).
The job of the ‘retail trader telco’ is performing that characterisation function, whereas the
‘wholesale trading platform’ executes the trades. This is a new demand-led ‘digital telco’
business model. It contrasts with the traditional supply-led model we are all familiar with,
with ‘one size fits all’ circuit and broadband products.
7. 7
A demand-led model is
a different kind of business
In a demand-led this model, the nature of the customer relationship evolves. The
retail/services telco function helps customers to describe their demand in terms that reflect
the performance levels on offer. This is more of a customer intimacy play than that of a
traditional scale-led telco, or the lock-in from vertical integration. More profit comes from
better characterisation of the demand for QoE (and performance) than the competition.
The wholesale telco function manages the ‘QoE portfolio’ and creates mechanisms that
support the most attractive set of ‘trading options’. Profit comes from a ‘platform fee’ for
executing trades. In our proposed prototype business unit, we would create a new wholesale
function to experiment with which retail channels (internal/external) fit a demand-led model.
This is more like an Uber or Airbnb than a traditional telco. There are still underlying network
assets needed, but like Uber and Airbnb, you don’t need to own them. This opens up the
opportunity to create pure ‘software company’ revenues by arbitraging your rivals’ asset
investments.
8. 8
Examples of possible new
‘quality arbitrage’ products
Examples of possible new products include:
• A range of resiliency service levels exploiting operationally-stranded assets
• Bulk data delivery service to increase intensity of use of assets
• Work with customers to optimise packet arrival patterns to reduce risk of QoE/cost shocks
• Quality arbitrage of 3rd party networks
• Bonded “hybrid access” services (increased capacity by quality-bonding multiple paths)
• Upstream assured access for closed user groups (key service protection)
• Downstream assured services delivered to user
• Assured backhaul over commodity broadband
• Home worker services
• Range extension of DSL lines
• Quality-assured cloud connectivity (capacity + quality trading optimisation)
This list is not exhaustive, and all would be subject to feasibility studies.
9. 9
Why create a new quality
arbitrage business unit and
product model?
10. 10
Why act now?
Immediate valuable learning
Telcos are at present ‘flying blind’ with respect to this arbitrage business risk. It is therefore
important to quantify both the risk and opportunity before the arbitrage becomes known and
widely exploited. By engaging with the initial scoping and validation phase of a new business
model you initiate a process of assessing the QoE risk of (part of) your product portfolio.
By the nature of being an arbitrage business, there is a first-mover advantage. Furthermore,
by doing the initial measurement process, you gather the information you need to find the
‘low-hanging fruit’ in terms of exploiting the arbitrage to your advantage. The sooner you
act, the better position you will be in to create a fundamental and hard-to-replicate strategic
differentiator from your competitors. SD-WAN is already extracting the arbitrage for users.
Political initiatives such as ‘net neutrality’ exacerbate the arbitrage risk. They increase the
danger of arbitrage, whilst constraining your ability to act. The learning from this project is a
useful piece of technical and economic counter-evidence to ‘neutrality fundamentalists’ in
the socio-political debate.
11. 11
Why is this the right thing to do?
Reposition yourself in the supply chain (1/2)
Managing the ‘trades’ of the arbitrage play form part of a wider ‘resource trading platform’. In
our opinion, the likely future profit generator of telecoms will be these platforms (e.g. SDN
orchestration), and not the underlying (bottleneck) physical assets. This is a very disruptive
‘value network orchestrator’ business model that repositions you in the supply chain.
Customers with high-intensity traffic between specific locations are currently motivated to
invest in bespoke capacity and to manage the resources for themselves. This process risks a
loss of your key enterprise and public sector ‘anchor tenants’. It can be reversed by
engaging with this demand-led business transformation.
By creating disruptive demand-led products, you can offer more value to customers by
delivering assured application and business outcomes. This differentiation drives revenue
growth.
12. 12
Why is this the right thing to do?
Reposition yourself in the supply chain (2/2)
By constructing disruptive new products, you offer convincing exemplars of success for
mainstream organisational adoption. Only a new business unit:
• Offers clear visibility to senior management of the comparative performance of the
business models.
• Cuts across current silo structure, creating a way for the functions to interact and create
new models and processes.
• Creates an institutionalised disruption process with feedback (that is currently missing).
A new business unit creates an incubator for developing new skills for dissemination
throughout the business:
• Establishment of the nucleus of network performance science skills
• Expertise in the new skill of managing ‘portfolio of QoE risk’
• Capability to run networks in saturation (for maximum returns) and measure QoE risk
13. 13
Why nothing else?
Only viable route to ‘move the needle’ (1/2)
We believe there are no politically acceptable defences against this arbitrage, e.g. a
regulation to ‘enforce a monopoly on multiplexing traffic’ is just not going to happen. The
only technical defence is to stop offering packet data transfer products!
Building a disruptive demand-led model requires new forms of interaction, both internal and
external, both of which demand a new business unit.
The internal mainstream organisation is likely to reject the vision, as it is at odds with its
current mission as a pure ‘economies of scale’ business. The disruptive products do not fit the
metrics currently uses to evaluate organisational and customer ‘success’.
Furthermore, large changes like this threaten to cannibalise its revenues and undermine its
social structure. These factors make it impossible for the existing structure to extract the
value of quality arbitrage.
14. 14
Why nothing else?
Only viable route to ‘move the needle’ (2/2)
Changing to a quality-centric focus requires a different mind-set, which is a culture-shock
issue in our experience.
Looking externally, a demand-led model requires a different form of customer relationship.
The retail telco’s job is to partner with users to deliver their desired application outcomes,
rather than selling them bandwidth. This is a customer intimacy play, not just traditional telco
scale.
Another new skill is the ability to translate these requirements into the contracts for
‘quantities of quality’ with the underlying wholesale platform.
There is a political and technical risk of executing an ‘attack’ arbitrage play against
competitors. A new unit could operate under a different brand, reducing reputational risk to
the existing business.
16. 16Monoservice vs Polyservice networks
A monoservice network offers essentially the same service to all packet flows (in bearer
terms, e.g. ‘share of bandwidth’). Such a service must satisfy the most stringent requirements
of any flow for low delay, low loss, and sequential delivery. In periods of low load any traffic
flow can experience the best possible service (as all classes have the same upper quality
bound).
A polyservice network is one in which different flows can be caused to experience
significantly different service. The quality of these services has bounds that differ between
classes. In application-relevant (rather than bearer) terms that means different levels of delay,
loss and/or sequentiality.
A network with multiple classes of service is a ‘monoservice network with QoS’. This QoS can
delivery differential levels of service when the network is stressed. The lower bound on the
quality of different classes is therefore different. However, at other times it makes no
difference, and therefore the upper bound on quality is the same for all classes.
17. 17Monoservice vs Polyservice summary
Lower quality* bound Upper quality* bound
Monoservice
Set to the most stringent
requirement of all flows (i.e.
maximum cost).
Highest possible, and same for all
flows
Polyservice
Variable and can be set with
assured bounds.
Variable and can be set with assured
bounds.
Monoservice
with QoS
Variable, but cannot be assured. Highest possible, and same for all
flows.
* Where ‘quality’ is the requirement on probability of delay, loss and non-sequentiality.
18. 18A polyservice network exploits the arbitrage
Monoservice
network
Polyservice
network
Efficiency
Effectiveness
Good
ΔQ
Poor
ΔQ
High
cost
Low
cost EfficiencyEffectiveness
Good
ΔQ
Poor
ΔQ
High
cost
Low
cost
Hard to improve
‘state of play’
Economy class
for bulk data
Superior class
for real-time
Standard class
for interactive use
19. 19
Over-delivery does not
guarantee consistent QoE
Quality
Time
Service A
quality failure
= churn driver
Service B excess
delivery = unnecessary cost
Service A
excess
delivery =
unnecessary cost
Quality delivered
to all services
Service A requirement
Service B requirement
20. 20
Polyservice delivers consistent
performance with low over-delivery
Quality
Time
Quality delivered
to service A
Quality delivered
to service B
Service A requirement
Service B requirement
21. 21
June 2009 - V 0.4
Monoservice
Polyservice
90% of load has consistent quality sufficient to get
TDM-like revenue; remainder has residual value
The result? The arbitrage is exploited,
more predictable performance, and lower cost
22. 22
Without deviation from the norm,
progress is not possible.
―Frank Zappa
Martin Geddes
mail@martingeddes.com