Thursday, October 25, 2012

The Opex / Capex Benefit Debate on Cloud

 Introduction

With the cloud space getting crowded each day, cloud service providers list a host of benefits by moving to the ‘Pay as you Go’ model…. Among other benefits, reduced capital expenses (fixed Capex costs), and increased operating expenses (variable Opex costs) is listed as a key benefit.
The article attempts to analyse this.
In general, from an accounting perspective, by shifting from a capex to opex based model, the benefits can be many.
Some of these could be:
-          Use of Opex allows deducting expense in the current year and reduces tax liability applied on net income.
-          Opex benefits due to time value of money (100 $$ saved today is better than 100 $$ saved next year)
-          No upfront capex expenditure if cost of capital is high and /or capital is not available
-          By Diverting  capex to opex, the capital saved can be used elsewhere giving a better yield.

All points taken, if you ask a accountant, he would say that it’s just accounting mechanics… He will also list down couple of points that go against this and as one can notice most of the cloud vendors do not mention this.
-          Opex decreases the company’s book value as the expense moves to profit & loss statement hitting the bottom line.
-          Opex decreases the company’s EBITDA.
-          Cost of On-demand services is always high as they carry a  utility premium. Think of renting a car vs. a buying a car. By renting a service, you are in a way borrowing money from the service provider, and paying the loan back to them in monthly payments. 
-          P&L, Balancesheet and cash flow statement looks less attractive.
More often than not, the real tangible benefits of moving from capex to opex are governed by other technical / business factors . Capex to opex movement is just an accounting measure and often tactical. It may well serve some short term financial objectives that the company needs but it is hard to prove that this is indeed true in the long run.
 What is Better : Capex or Opex
The
Should one really care about the opex and capex debate and what should be the choice? Or this is contextual. Often, the reasons of moving from Capex to Opex could be very different and purely financial. For e.g:
1)     Shortage / Availability of capital,
2)     High Cost of Capital,
3)     Overall economic environment which impact investment decisions.
4)     Better use of available capital
5)     Benefits from alternate capex investment outweighs the higher cost of ‘pay as you go service’ in the long run.
The objectives could be the other way around also (Reduce Opex and increase Capex), for e.g,:
-          To boost profits (artificially, some may say) to report to investors and increase shareholder value
-          A higher value of assets on its balance sheet
-          Cost of capital is cheap and readily available.
-          No better use of capital
What this means is that the movement of capex to opex or vise versa is highly contextual.  It may all sound good due to current economic environment and capital constraints but may  not hold good in all cases and in the long run
 The Real Benefits
The real benefit must be seen from both Financial as well as business / technical objectives. Once the business / technical objectives are ascertained, the appropriate financial treatment (capex / opex) need to be done based on the objectives / constraints. For example, the financial needs of a company could be
1)     Optimizing capital investment due to shortage / Availability of capital
2)     Optimize overall Cost of Capital
3)     Workaround the difficult economic environment which impact investment decisions.
4)     Investment Options at hand
5)     ROI
6)     Better profit reporting to investors and increase shareholder value
7)     Higher value of assets on balance sheet
8)     Better Cash flow management
The financial objective may be tactical in many cases and may change over a course of time.
That said, the real reasons behind a ‘pay as you go’ model or so called ‘opex’ model, may be unrelated to the financial context with very different objectives (Often business driven). These could be:
-          Better Utilization Rates
Here is where the real benefits of shared services come into play.  If a company is using less than 30% of installed capacity, ‘pay as you go’ model can give significant cost benefits. If it is already using more than 70 %, the benefits may not be that high as utility services come at a premium as compared to owned service for obvious reasons..

-          Operational advantages

-          Service level improvements

-          User satisfaction considerations

-          Keeping pace with the technology upgrades

Due to the rapid changes in technology,  it doesn’t make sense to sink or locking money into equipment that’s surpassed by the very next model. Opex in the form of pay as you go models help in this regard. Costs are reduced and shifted from capital expenses to operating expenses, which save money. The capital saved can be used elsewhere in more productive investments or reduce the cost of capital.

-          Better scalability, metering, automation and virtualization – all for less money than a traditional data center.  

            Bottom line is that CAPEX to OPEX shift should rarely be the deciding factor in a systems selection.
            The choice architecture should be based upon meeting the business and technical requirements.

A company can choose to use a public shared cloud - on rent and hence opex based model OR go for a private cloud  - owned and capex OR an internal cloud (virtualized) that can satisfy all the business / technical requirements for the CIO while also meeting the financial objectives for the CFO.

It all depends on the context. A WIN-WIN situation for all is one that should be looked at.


________________________________________________________________________________

Brief on Opex and Capex

Capex
-          Capital expenditures are expenses to create assets (or add value to an existing asset) for future benefts with a life that extends beyond the tax year.
-          Cannot be fully deducted in the period when they were incurred. Tangible assets (building) intangible assets ( h Money spent on inventory falls under capex ardware etc) are depreciated and intangible assets (patent etc) are amortized over time.
-          Money spent on inventory falls under capex
-          Show up in Balance Sheet and Cash Flow from Investing head in the Cash flow statement

Opex
         -          Refers refers to expenses incurred in the course of ordinary business, such as sales, general and    
            administrative expenses (and excluding COGS, taxes, depreciation and interest
       -     Operating expenses are fully deducted in the accounting period during which they were incurred.
       -         Show up in Profit and loss statement.
       -          End up in Cash Flow from operating activities head in the Cash flow statement

Sunday, October 7, 2012

Cloud Pricing Models

Analysis, Comparison and Trends for IaaS and SaaS vendors

Introduction
The current pricing of Cloud based services offers a wide range of configuration, choices and price points for the users. This, in turn creates lot of choice (and confusion) to the buyer as it is difficult to measure the value. 
The objective of this article is to analyze the existing pricing models and reasons behind the pricing diversity. In the following 4 sections, this article attempts to:

1)    Compare approx. 27 models of pricing derived from the scholarly article
2)    Recommend a derived pricing model relevant to the cloud service.
3)    Validate the model by mapping it to the pricing of services by few key vendors
4)    Analyze reasons behind the pricing in use by various vendors and future trends.

1. Generic Pricing Models
  • Osterwalder (2004) classified the pricing models around fixed, differential and market based pricing. He suggested that the fixed and differential pricing mechanisms produce prices that depend on customer/ product / Service characteristics, volume, but are not based on real-time market conditions. Market based pricing stands for pricing mechanisms that produce prices based on real-time market conditions.
  • Harmon et al. (2009) suggested a factor based pricing mechanism (cost + value).
  • Denne (2007) discussed various advanced ways to implement pay per use pricing mechanism.
  • Paleologo (2004) suggested that traditional pricing mechanisms such as cost-plus pricing may be inadequate in on-demand services environment due to dynamic factors in cloud computing as shorter contracts, reduced switching costs, customer lock-in, uncertain demand, and shorter life cycles etc.
  • Some authors have also suggested other ways of classification such as resource based Vs. Feature based or Operational (Pay per Use) Vs. Availability (subscription).

2. A Derived Pricing Model for Cloud Computing 

From this article’s perspective, we classify the pricing models which derived from the discussed pricing mechanisms.


  1. Usage Based Pricing (Pay per use) with factor based variations.
  2. Subscription Pricing (Commitment / Reserve / Availability Pricing) with factor based variations.
  3. Market Based Pricing
  4. Pricing Strategy (More relevant from a marketing perspective)

Following table summarizes the derived pricing model:



3. Validation of the Derived Pricing Model

The table below provides a mapping of the derived pricing model to the pricing of 9 cloud offerings by few key vendors.































































4. Analysis and Future Trends  

  • Pricing matrix is more complex for IaaS compared to SaaS as service offering can be better differentiated from an end user perspective.
  • Within IaaS, compute pricing has more variety compared to storage or bandwidth. The reason for this variation is the fact that storage and bandwidth offer very less flexibility in terms of configuration as compared to compute service offering which is used to create differentiation.
  • In compute services, companies are trying to offer pseudo differentiation by packaging / configuring the service features in various ways. This is leading to more confusion for the buyers as it difficult to measure and value the offering in a meaningful way. Measuring and comparing CPU, memory, speed etc. is difficult for an offering and to also to make a decision since the compute instance is a vendor specific configuration in most cases.  
  • As the market matures and buyers eventually become smarter, the pricing of the compute services will become smarter and simpler for the buyer. The differentiation will be done on more tangible parameters such as guaranteed SLA, support, processor speed, performance etc. rather than the configuration and packaging which look almost the same (or ‘different’ from the vendor’s perspective).
  • The trend for compute services at the moment is lead by Amazon (being the first movers) and all others following them with  more or less same kind of pricing and configuration models. They were the first to bring market based pricing (spot instances) which is yet to be copied by other vendors. This trend should continue.
  • As services become commoditized, service brokers / exchanges or ecosystems will emerge (many already in place as jam cracker, thinkgrid etc). This will be the starting point of more sophisticated market based complex pricing models for the exchanges and the brokers. The complexity will however be shielded from the end customer / buyer. Pricing definition will shift from the service provider to the marketplace requirements. Outside the market place, companies will deploy more established pricing strategies to gain market share including competitive pricing, discount schemes, customer loyalty, volume based discounts, yield management the etc. which is mainly the usp of enterprise services at the moment.
  • SaaS market is highly feature / functionality based and pricing is simpler and easier for the user as they can easily relate the value to the features / functionality. As the competition increases, companies will deploy new pricing strategies to gain market share including competitive pricing, discount schemes, customer loyalty bonuses, volume based discounts etc.
  • PaaS needs a different treatment and is still in infancy stage.
  • With multiple providers competing to deliver very similar commoditized services in a high price sensitive environment, the market will able soon approach a perfect competition like situation and the “Me Too” type of offering is not going to work. This will trigger : 

         a) Price war (And possibly the “Race to Bottom”) unless the vendors leverage  
             economies of scale, shared infrastructures, reduced deployment costs, and free and 
             open source versions etc to beat the competition. 
         
         b) Value Added / differentiated offering that can be tangibly defined, measured and can 
              be  paid in terms of support, maintenance, SLAs, and performance etc. will charge a 
              premium.
 


References

[1] Osterwalder, A. (2004). The Business Model Ontology - A Proposition In A Design Science Approach. Doctoral thesis. University of Lausanne.

[2]   Denne, M. (2007). Pricing utility computing services. International Journal of Web Services Research, Vol. 4, No. 2, pp. 114—127.

[3]  Harmon, R., Demirkan, H., Hefley, B. & Auseklies, N. (2009). Pricing Strategies for Information Technology Services: A Value-Based Approach. Proceedings of 42nd Hawaii International Conference on System Sciences, pp. 1—10.

[4]  Paleologo, G. (2004). Price-at-Risk: A methodology for pricing utility computing services. IBM Systems Journal, Vol. 43, No. 1, pp. 20—31


About Me

    

Sunil has over 15 years experience in Information Technology providing leadership, management, planning, system development and engineering, training, people development, methodologies, and process support. He has worked in the area of internet based technology solutions, e-commerce applications, product development, product maintenance, e-business platforms, portals, collaboration, content management, and business intelligence. He currently works for Colt Technologies Services, Bangalore, India as Head of Systems Development and Support. Sunil holds an Executive MBA (IIM Calcutta), Masters in Technology (IIT Kanpur) and a Bachelor of Engineering in Electrical Engineering. He also holds a diploma in entrepreneurship from EDI Ahmedabad.

He can be contacted at 
Linkedin: http://www.linkedin.com/pub/sunil-pathak/4/501/a9b
Twitter.com/sunil__pathak