In the closing months of last year we conducted a survey of our clients as to their views on the benefits and risks of cloud computing.  Admittedly, conducting surveys of IT professionals and business executives regarding their views on cloud computing has been somewhat of a growth industry over the past two years.  However, our survey was notably different in that most of the respondents were either in-house counsel or in risk management positions wherein they are responsible for retaining counsel, rather than business executives or IT professionals.  As a result, the survey has provided a valuable cross-section of those critical personnel's views as to the risks and benefits of cloud computing.

A. The Questions

The questions effectively fell into four categories:

  1. Extent of knowledge:  How would you express your current understanding of cloud computing?
  2. Perception of risks:  How would you rank in importance the risks regarding cloud computing?  Do you perceive these to be "new" risks - e.g. as compared to a traditional hosting arrangement?
  3. Perception of benefits:  How would you rank in importance the benefits of cloud computing?
  4. Actual experience:  Have you adopted any cloud computing models in your organization?  If yes, what was your experience?  If no, for what reasons did you reject this alternative delivery mechanism?

B. The Responses

The survey results are outlined below:

1. Extent of Knowledge

Approximately 75% of the respondents indicated that they had an understanding of cloud computing which was "basic" to "good", with about 25% responding that they had "poor" or "no" understanding, a result which is generally consistent with other surveys in the sector.[1]

2. Perception of Risks

We asked our respondents how each would rank the following concerns regarding cloud computing.  The top four concerns were:

  • incompatibility with existing agreements, with 64%, or more than 2/3, of the respondents indicating that they had a high level of concern with the issue of legal compatibility, and 26% ranking this concern as very high;
  • inability to customize standardized cloud service offerings, which would also relate to the inability to customize agreements, and which tied with legal compatibility as having 64% of respondents ranking such inability as a high to very high level of concern;
  • lack of knowledge as to risks/benefits, which 51% of the respondents ranked as high, with 30% ranking the concern as very high; and
  • having data returned in a timely manner when required (e.g. in a litigation context), which tied with the lack of knowledge concern as being the second greatest concern. 

Three other concerns which were also identified as being material were:

  • lack of visibility as to location of data, such that 42% ranked lack of visibility as a high level of concern, and 11% ranked the issue as very high;
  • inability to satisfactorily negotiate one-sided vendor terms and conditions, wherein 36% of the respondents ranked this concern at a high to very high level; and
  • insufficient vendor service commitments, wherein whilealmost 1/3 of the respondents ranked this at a high level of concern, approximately another 1/3 identified this of being at a low level of concern only

The three lowest ranked concerns were as follows:

  • privacy concerns -while approximately ¼ of the respondents identified these issues as being of a high level of concern, the majority of the respondents (2/3) noted that the issue was of a low concern only;
  • compliance with regulatory requirements re: location of data, where again, approximately 2/3 of the respondents ranked regulatory compliance as being their lowest concern; and
  • security concerns, which - for our respondents - was the issue of least concern, with only 9% ranking security as being of "high" concern, and a clear majority - 76% - ranking security as being of a low level of concern.

What is interesting about these results is the significant divergence between these responses and the responses of other surveys - again, of business executives and IT professionals - to the same questions.  In contrast to our survey, other surveys have consistently identified security as being the primary concern[2]; privacy as being a key concern[3]; and regulatory compliance as being a top concern,[4] while - according to one CIO survey - "performance issues, contracts and difficulties around service level agreements are all moderate barriers, but never cited by more than 20% of all respondents"[5]

In other words, when business executive and IT professionals are surveyed, their list of concerns is almost the inverse of those of our respondents from the legal/risk management community[6].

Finally, in answer to the question as to whether the respondents perceived these to be "new" risks - e.g. as compared to a traditional hosting arrangement - 58% answered "yes", while 36% answered "no".  While prior incarnations of related models, whether hosting arrangements, virtualization models or SaaS, suggest that this is not in fact the case, that is a topic to be addressed in another bulletin.[7]

3. Perception of Benefits

We also asked each of our respondents to rank the benefits of cloud computing.  The top three benefits were as follows:

  • improved performance, such that40% ranked the benefit of improved performance at a high level, and 23% ranked the benefit as being the most important benefit;
  • standardization, with 40% of our respondents ranking this benefit at a high level, and 19% ranking standardization as the most significant benefit; and
  • utility pricing, with 32% ranking the benefit at a high level, and 15% ranking utility pricing as the number one benefit.

The following were the lowest ranked benefits:

  • lower costs were one of the lowest ranked benefits:  surprisingly, while 28% ranked the benefit of lower costs at a high level, with 21% ranking this benefit as the most significant, more of the the respondents (34%) ranked the benefit of lower costs as being at a low priority.
  • device and location independence, so while 28% ranked the benefit of device and location independence at a high level, more 38% ranked the benefit of device and location independence at a low level, and 17% ranked the benefit of device and location independence as the least significant benefit;
  • scalability, with 34% ranking the benefit of scalability at a low level, and 8% ranking the benefit of scalability at the lowest ranking; and
  • convenience (e.g. turn-key) was ranked by 40% at a low level, with 21% ranking the benefit of convenience as the lowest priority.

How do these results compare to other cloud computing surveys?  Other surveys' responses have identified the most significant benefits as being:

  • the most cited business drivers were lowering the cost of ongoing IT operations (54%)  and reducing capital  investment in servers and data center equipment (51%) (US Federal Government Survey 2013[8]);
  • top benefits noted were cost savings (65%), flexibility/scalability (58%), and efficient use of resources (CA Technologies Survey[9]);
  • the biggest drivers were agility, scalability and cost savings - 55% of the respondents believe that that the cloud has a lower Total Cost of Ownership (451 Group Survey[10]); and
  • scalability/flexibility (30%) and reduced hardware infrastructure costs (28%):  the CIO Survey.

Again, we believe that the risk management role of our respondents at least partially explains the less technical orientation of their responses.

4. Actual Experience

We then asked our respondents if their respective companies had adopted cloud computing for any services, and if they responded positively, what were their experiences.  In response, a narrow majority of 53% indicated that they had adopted cloud computing for certain services (the "cloud adopters"), leaving 43% who indicated that they had not (the "cloud rejecters"). 

Of those who responded positively:

  • with respect to question as to which departments had adopted cloud computing, 43% indicated IT, 21% responded HR, and marketing and finance tied at 7%; however, 25% responded "Other";
  • more than one third (39%) indicated that they had adopted a "public" model, whereas 29% confirmed adoption of a "private cloud" model, 18% confirmed a "community" model and 14% answered "hybrid" model.
  • the cloud adopters reported that the primary reasons for adopting a cloud computing model were convenience (43%) and lower costs (29%); 14% responded with "device and location independence"; and less than 10% in each case responded with "scalability", "standardization" and "improved performance", results which were more consistent with other cloud computing surveys - in making sense of these results, it appears that the benefits which actual cloud adopters found persuasive, were different from those contemplated by other groups like the cloud rejecters; and
  • recognizing that there are a number of strategies available to mitigate the risks of cloud computing, the most popular strategy of cloud adopters (36%) was to "only cloud less sensitive data".  The next most popular strategies were the successful negotiation of protections in terms and conditions (18%) and adopting a hybrid model (14%).  Less popular responses included "increased internal due diligence re whether permitted by licences, etc." (11%) and "increased external due diligence as to data location" (7%).

Finally, when surveyed regarding their overall experience with the cloud, 86% of the cloud adopters responded that it was excellent or good.  However, when asked to rank their experience in negotiating the cloud computing agreements with the vendor, the results were more evenly split:  47% responded that their experience had been poor or very poor, 50% responded that it had been good, and no one responded that it had been excellent.

We also asked each respondent if their your company had ever considered, and rejected, adopting a cloud computing model, to which 58% answered "no", but 38% answered "yes".  Of the cloud rejecters:

  • Two responses tied, at 25%, for the primary reason for rejecting the cloud model:  (a) inability to satisfactorily negotiate one-sided vendor terms and conditions, and (b) security concerns.  Next ranked were "privacy concerns" (20%), and "compliance with regulatory requirements re location of data" (10%).
  • The reasons which were cited least (all at 5%) were:  "lack of visibility as to location of data"; "insufficient vendor service level commitments"; concern regarding having data returned in a timely manner when required (e.g. in a litigation context); and "incompatibility with requirements of existing license/third party agreements".

Note that these concerns of this subset of respondents - that is, of the "cloud rejecters" - are more closely aligned with other surveys of the sector:  i.e. with security concerns, privacy concerns and compliance concerns ranking highly. 

We can theorize that the reason for the distinction may be that the "cloud adopters", having taken the plunge and implemented one or more cloud models, now perceive the risks more differently than those cloud rejecters which did not proceed with a cloud model, perhaps because of their lived experience:  recall that 86% of cloud adopters responded that their actual experience with the cloud was excellent or good, once implemented.

The Conclusion

It is interesting to note that both cloud adopters and cloud rejecters agreed that their experience in negotiating the terms and conditions for the cloud arrangement was generally very poor.  For cloud rejecters, it tied with security concerns as being the primary reason for rejecting the cloud model.  Even for cloud adopters, almost half (47%) responded that the negotiating experience was either poor or very poor.

This is consistent with not only our experience, but also the experience of others in the sector.  The recent Backbone article "Why large Canadian companies don't want anything to do with cloud computing", in reviewing how "adoption of cloud computing among Canada's larger firms is surprisingly low", with "71% of large Canadian enterprises not open to alternative IT delivery models", noted that "some cloud providers, such as Freshbooks, don't provide service level agreements (SLA's) because they are catering to small business customers with non-mission-critical services".[11]

Other articles, however, suggest that the problem is more wide-spread than for just small and medium size businesses ("SMB's").   A recent paper, "Negotiating Cloud Contracts - Looking at Clouds from Both Sides Now", reported that:

The starting point for cloud contracts is almost always the providers' standard terms and conditions, which unsurprisingly  are provider-favorable and are generally designed for high  volume,  low  cost,  standard,  commoditized  services  on  shared  multi-tenant infrastructure.  However,  as  many  providers'  standard  terms  are  generally  not  apt  to accommodate enterprise users' requirements, cloud users had sought changes to make the  terms  more  balanced  and  appropriate  to their  own  circumstances. It  appears that there has been some movement, particularly for large users. Nevertheless, our research indicates that some providers' negotiations  are still very process-driven, particularly at the  lower  price  end  of  the  market,  where  providers  seemed  unable  or  unwilling  to accommodate  differences  such  as  corporate  structures  entailing  (for  users)  separate localized contracts for non-US affiliates… [12]

Our findings tie in with a survey for CIF of 450 senior IT and business decision makers in  enterprises,  SMEs  and  public  sector  organizations,  published  in  early  2011.  This found  that,  while  48%  of  organizations  polled  were  already  using  cloud,  only  52%  of those (particularly larger organizations) had negotiated their contracts, with 45% stating that they had no opportunity to negotiate (click-through terms, discussed below). In other words,  according  to  that  survey,  about  half  of  such  users'  cloud  contracts  were negotiated.

However, the authors conclude on an optimistic note:

Factors affecting development of cloud contract terms will be both demand and supply led.  More  customer-friendly  terms  are  being  demanded  by  large  users  such  as  government,  and  being  offered  or  agreed  by  integrators,  smaller  or  niche/specialist providers  and  market  entrants,  thus  making  contract  terms  a  source  of  competitive advantage  for  providers.  As such,  we  believe  that  standard  terms  will  improve  from users' perspective as the market develops further. [13]

We can only hope that, as we start this New Year, providers will increasingly recognize the significance of one-sided, take-it-or-leave contractual terms as an impediment to the broader adoption of cloud computing models.