Coût d’une perte de données en France (Livre Blanc en anglais)


Published on

Pour la 4ème année consécutive, le Ponemon Institute publie les résultats de son étude menée auprès d’entreprises françaises sur les coûts liés à une perte de données, en fonction du type d’attaquant, des secteurs d’activité, des structures organisationnelles…

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Coût d’une perte de données en France (Livre Blanc en anglais)

  1. 1. 2013 Cost of Data Breach Study:FranceBenchmark research sponsored by SymantecIndependently Conducted by Ponemon Institute LLCMay 2013Ponemon Institute© Research Report
  2. 2. Ponemon Institute© Research Report Page 120131Cost of Data Breach Study: FrancePonemon Institute, May 2013Part 1. Executive SummarySymantec Corporation and Ponemon Institute are pleased to present the 2013 Cost of DataBreach: France, our fourth annual benchmark study concerning the cost of data breach incidentsfor companies located in France. For organizations in France, the cost of a data breach continuesto rise. In 2012 the cost increased from !122 to !127 on a per capita basis.2Ponemon Institute conducted its first Cost of Data Breach study in the United States eight yearsago and France four years ago. Since then, we have expanded the study to include Germany,United Kingdom, Italy, India, Japan and Australia and, for the first time this year, Brazil. To date,87 French organizations have participated in the benchmarking process since the inception of thisresearch four years ago.Since Ponemon Institute began studying this issue, several EU countries have enacted lawsrequiring the controller of databases that contain personal information to inform affectedindividuals in the event of data loss or theft. In an effort to reduce administrative burdens and thecost of compliance with data protection laws, including data breach notification, the EuropeanCommission announced a proposal to reform the European Union’s data protection framework.Announced in January 2012, the proposed regulation creates a single set of European rules thatwould be valid everywhere for all EU member countries.3This year’s study examines the costs incurred by 26 French companies from 10 different industrysectors following the loss or theft of protected personal data and notification of data breachvictims as required by various laws. It is important to note the costs presented in this research arenot hypothetical but are from actual data loss incidents. The costs are based upon estimatesprovided by individuals interviewed over a ten-month period in the companies represented in thisresearch.To calculate the cost of data breach, the research studies a wide range of business costs,including expense outlays for detection, escalation, notification and after-the-fact (ex-post)response activities. We also analyze the economic impact of lost or diminished customer trustand confidence as measured by customer turnover or churn rates. The number of breachedrecords per incident this year ranged from 2,381 to 72,186 and the average number of breachedrecords was 22,462. We do not include organizations that had data breaches in excess of100,000 because they are not representative of most data breaches and to include them in thestudy would skew the results. The cost for the 26 data breach case studies in this year’s report ispresented in Appendix 1.The following are the most interesting findings and implications for organizations:! The cost of data breach increased. For the fourth consecutive year, the cost per lost orstolen record and the total organizational cost have increased. The average cost of databreach rose from !122 in 2011 to !127 in 2012. We define a record as information thatidentifies an individual and regulations require notification of data breach victims.The average total organizational cost of data breach increased from !2.55 million to !2.86million – or, an 11 percent increase between 2011 and 2012 results. This increase suggeststhe need for organizations to improve their ability to respond to the breach.1The Cost of Data Breach report is dated as a 2013 publication. Please note that all data breach incidents studied in thisyear’s report happened in the 2012 calendar year. Thus, all figures reflect the 2012 data breach incidents.2The terms “cost per compromised record” and “per capita cost” have equivalent meaning in this report.3See: European Commission Proposal for a Regulation of the European Parliament and of the Council on the Protectionof Individuals with Regard to the Processing of Personal Data and on the Free Movement of Such
  3. 3. Ponemon Institute© Research Report Page 2! Customer churn rates continue to be high. Abnormal customer churn or turnover (a higherthan average loss of customers for the industry or organization) in this year’s study stayedconstant from last year at 4.4 percent on average. Certain industries, such as financialinstitutions and service companies, are more susceptible to customer churn, which causestheir data breach costs to be higher than the average. Our research suggests taking steps tokeep customers loyal and repair any damage to reputation and brand can help reduce thecost of a data breach.! Malicious or criminal attacks are most often the root cause of the data breach. Forty-two percent of organizations say the root cause involved the theft or misuse of information,including exfiltration by external attackers. Thirty-one percent of breaches involved negligentemployees or contractors (a.k.a. human factor) and 27 percent say it was due to IT andbusiness process failures. Accordingly, organizations need to focus on processes, policiesand technologies that address threats from the malicious insider or hacker.The average per capita cost of data breach relating to the theft or exfiltration of data is !142.In contrast, the average cost of data breaches relating to system glitches and human errorsare !118 and !116, respectively.! Lost business costs increased sharply from !.78 million in 2011 to !1.19 million in2012. This is the highest lost business cost reported since the first study in 2009. These costsrefer to abnormal turnover of customers, increased customer acquisition activities, reputationlosses and diminished goodwill.! Certain organizational factors decrease the overall cost. Organizations that had arelatively strong security posture at the time of the data breach saved as much as !12 percompromised record. Organizations that had an incident response plan in place before thedata breach incident realized cost savings of as much as !7 and those that appointed a CISOwith overall responsibility for enterprise data protection experienced savings of !4 percompromised record. Outside consultants assisting with the breach response saved !5 perrecord. When considering the average number of records lost or stolen, these factors canprovide significant and positive financial benefits.! Certain organizational factors increase the overall cost. Specific attributes or factors ofthe data breach also can increase the overall cost. For organizations in this study, databreaches involving third parties increased the cost by !17 per compromised record. Inaddition, data breach incidents involving the loss or theft of a computer or storage deviceincreased the cost by !7 per compromised record. In addition, quick notification increased theper capita cost by !4.! Ex-post response and detection costs decreased slightly. The costs associated with ex-post response decreased from approximately !.91 million in 2011 to !.83 million in 2012. Ex-post response costs refer to all activities that attempt to address victim, regulator and plaintiffcounsels’ concerns about the breach incident. This cost category also includes legal andconsulting fees that attempt to reduce business risk and liability. Redress, identity protectionservices and free or discounted products are also included in this cost category.Similarly, the costs associated with detection and escalation activities decreased from !.75thousand in 2011 to !.72 thousand in 2012. This category refers to activities that enable acompany to detect the breach and determine its root cause. It also includes upstream andlateral communications that are required to focus activities on data breach resolution andkeep management informed.
  4. 4. Ponemon Institute© Research Report Page 3Cost of Data Breach FAQsHow do you collect the data?Ponemon Institute researchers collected in-depth qualitative data through interviews conductedover a ten-month period. Recruiting organizations for the 2012 study began in January 2012 andinterviews were completed in December. In each of the 26 participating organizations, we spokewith IT, compliance and information security practitioners who are knowledgeable about theirorganization’s data breach and the costs associated with resolving the breach. For privacypurposes we do not collect any organization-specific information.How do you calculate the cost of data breach?To calculate the average cost of data breach, we collect both the direct and indirect expensespaid by the organization. Direct expenses include engaging forensic experts, outsourced hotlinesupport, free credit monitoring subscriptions and discounts for future products and services.Indirect costs include in-house investigations and communication, as well as the extrapolatedvalue of customer loss resulting from turnover or diminished acquisition rates. For a detailedexplanation about Ponemon Institute’s benchmark methodology, please see Part 4 of this report.How does benchmark research differ from survey research? The unit of analysis in the Costof Data Breach study is the organization. In survey research, the unit of analysis is the individual.As discussed previously, we recruited 26 organizations to participate in this study.Can the average cost of data breach be used to calculate the financial consequences of amega breach such as those involving millions of lost or stolen records?The average cost of a data breach in our research does not apply to catastrophic breaches.Primarily because these are not typical of the breaches most organizations experience. In orderto be representative of the population of French organizations and draw conclusions from theresearch that can be useful in understanding costs when protected information is lost or stolen,we do not include data breaches of more than 100,000 compromised records.Are you tracking the same organizations each year?Each annual study involves a different sample of companies. In other words, we are not trackingthe same sample of companies over time. To be consistent, we recruit and match companies withsimilar characteristics such as the company’s industry, headcount, geographic footprint and sizeof data breach. Since starting this research in 2009, we have studied the data breachexperiences of 87 organizations in France.
  5. 5. Ponemon Institute© Research Report Page 4Part 2. Key FindingsIn this section we provide the detailed findings of this research. Topics are presented in thefollowing order:! Cost of data breach per record and organization! Cost of data breach by industry! Root causes of a data breach! Factors that influence the cost of a data breach! Trends in the frequency of compromised records! Trends in customer turnover or churn! Trends in the following cost components: detection and escalation, notification, lost business,direct and indirect and post-data breach! Preventive measures taken after the breach! Percentage changes in cost categoriesThe per record and organizational cost of data breach continues to increase. Figure 1reports the average per capita cost of data breach.4As can be seen, for four consecutive yearsthe average per capita cost has increased. This year, data breaches cost companies an averageof !127 per compromised record – of which !71 pertains to indirect costs. This includes abnormalturnover or churn of existing customers. Last year’s average per capita cost was !122 with anaverage indirect cost of !70.Figure 1. The average per capita cost of data breach over four yearsBracketed number defines the benchmark sample size4Per capita cost is defined as the total cost of data breach divided by the size of the data breach in terms of the number ofcompromised records.! 89! 98! 122! 127! 0! 20! 40! 60! 80! 100! 120! 1402009 (17) 2010 (21) 2011 (23) 2012 (26)
  6. 6. Ponemon Institute© Research Report Page 5Average organizational cost of data breach increased. Figure 2 shows that the total averagecost of data breach over four years has trended upward. In this year’s study, the totalorganizational cost of data breach increased from !2.55 million to !2.86 million – or, an 11percent increase between 2011 and 2012 results.Figure 2. The average total organizational cost of data breach over four yearsKey cost of data breach measures. Figure 3 reports the increases in per capita cost and theaverage total data breach cost from 2011. The average data breach size or number of recordsincreased by 7 percent and abnormal churn rate is virtually the same as last year’s result. In thecontext of this research, abnormal churn is defined as the greater than expected loss ofcustomers in the normal course of business.Figure 3. Cost of data breach measuresNet change defined as the difference between the 2012 and 2011 results! 1.90! 2.20! 2.55! 2.86! 0.00! 0.50! 1.00! 1.50! 2.00! 2.50! 3.00! 3.502009 (17) 2010 (21) 2011 (23) 2012 (26)0%4%7%11%0% 2% 4% 6% 8% 10% 12%Abnormal churnPer capita costAverage size of data breachAverage total cost
  7. 7. Ponemon Institute© Research Report Page 6Certain industries experience more costly data breaches. Figure 4 reports the per capitacosts for the 2012 study by industry classification. While a small sample size prevents us fromgeneralizing industry cost differences, the pattern of 2012 industry results is consistent with prioryears. As can be seen, consumer products, industrial, pharmaceuticals and financial serviceorganizations experience a much higher per capita cost than the sample mean. In contrast, publicsector, energy and retail companies experience a per capita cost substantially below the meanvalue.Figure 4. Per capita cost by industry classification of benchmarked companies! 81! 82! 82! 103! 140! 143! 146! 153! 154! 161! 0 ! 20 ! 40 ! 60 ! 80 ! 100 ! 120 ! 140 ! 160 ! 180Public sectorEnergyRetailTechnologyServicesHospitalityFinancialPharmaceuticalIndustrialConsumer
  8. 8. Ponemon Institute© Research Report Page 7Malicious or criminal attacks are the primary root causes of a data breach. Figure 5provides a summary of the main root causes of a data breach for all 26 organizations. Forty-twopercent experienced a malicious or criminal attack.5Thirty-one percent of incidents involved anegligent employee or contractor (human factor)6, and 27 percent involved system glitches,including a combination of both IT and business process failures.Figure 5. Distribution of the benchmark sample by root cause of the data breachMalicious attacks are most costly. Data thieves or disgruntled insiders were the most likelycauses of data breach as reported by our sample of 26 companies. Figure 6 reports the percapita cost of data breach for three conditions or root causes of the breach incident, includingmalicious attacks.Again, the pattern of results in 2012 is consistent with previous studies, where the most costlybreaches involved malicious acts against the company rather than negligence (human factor) orsystem glitches. Accordingly, companies that experienced malicious or criminal attacks had thehighest average per capita cost (!142), and companies that experienced system glitches had aper capita cost of !118. Negligence resulted in the lowest per capita cost of !116.Figure 6. Per capita cost for three root causes of the data breach5Malicious and criminal attacks decreased slightly from 43 percent in our 2011 study. The most common types of attacksinclude malware infections, criminal insiders, phishing/social engineering and SQL injection.6Negligent insiders are individuals who cause a data breach because of their carelessness, as determined in a post databreach investigation.42%27%31%Malicious or criminal attackSystem glitchHuman factor! 142! 118 ! 116! 0! 20! 40! 60! 80! 100! 120! 140! 160Malicious or criminal attack System glitch Human factor
  9. 9. Ponemon Institute© Research Report Page 8Seven factors that influence the cost of data breach. We identified seven factors thatinfluence the cost consequences of a data breach incident. These attributes are as follows:! The company had an incident management plan. Thirty-two percent of organizations inour benchmark sample had a data breach incident management plan in place at the time ofthe data breach event.! The company had a relatively strong security posture at the time of the incident.Thirty-nine percent of organizations had a security effectiveness score (SES) at or above thenormative average. We measured the security posture of each participating company usingthe Security Effective Score (SES) as part of the benchmarking process.7! CISO (or equivalent title) has overall responsibility for enterprise data protection.Thirty-five percent of organizations have centralized the management of data protection withthe appointment of a C-level information security professional.! Data was lost due to third party error. Forty-eight percent of organizations had a databreach caused by a third party, such as vendors, outsourcers, cloud providers and businesspartners! The company notified data breach victims quickly. Forty-two percent of organizationsnotified data breach victims within 30 days after the discovery of data loss or theft.! The data breach involved lost or stolen devices. Thirty-nine percent of organizations hada data breach as a result of a lost or stolen mobile device, which included laptops, desktops,smartphones, tablets, servers and USB drives containing confidential or sensitive information.! Consultants were engaged to help remediate the data breach. Twenty-nine percent oforganizations hired consultants to assist in their data breach response and remediation.As shown in Figure 7, strong security posture, incident response planning, CISO appointment andconsulting support decreased the per capita cost of data breach. In contrast, third party error,quick notification and lost or stolen devices increased per capita cost. Hence, a strong securityposture reduced the average cost of data breach from !127 to !115 (decreased cost = !12). Incontrast, a third party error increased the average cost to as much as !144 (increased cost =!17) per compromised record.Figure 7. Impact of seven factors on the per capita cost of data breach7The Security Effectiveness Score was developed by Ponemon Institute in its annual encryption trends survey to definethe security posture of responding organizations. The SES is derived from the rating of 24 security features or practices.This method has been validated from more than 40 independent studies conducted since June 2005. The SES provides arange of +2 (most favorable) to -2 (least favorable). Hence, a result greater than zero is viewed as net favorable.-! 12-! 7-! 5-! 4! 4! 7! 17-! 15 -! 10 -! 5 ! 0 ! 5 ! 10 ! 15 ! 20Strong security postureIncident response planConsultants engagedCISO appointmentQuick notificationLost or stolen devicesThird party errorDifference from mean
  10. 10. Ponemon Institute© Research Report Page 9The more records lost, the higher the cost of data breach. Figure 8 shows the relationshipbetween the total cost of data breach and the size of the incident for 23 benchmarked companiesin ascending order by the size of the breach incident. The regression line clearly indicates that thesize of the data breach incident and total costs are linearly related. In this year’s study, the costranged from !239,431 to !10,132,750.Figure 8. Total cost of data breach by size of lost or stolen recordsRegression = Intercept + {Size of Breach Event} x !, where ! denotes the slope.The more churn, the higher the per capita cost of data breach. Figure 9 reports thedistribution of per capita data breach costs in ascending rate of abnormal churn. The regressionline is upward sloping, which suggests that abnormal churn and per capita costs are linearlyrelated. This pattern of results is consistent with benchmark studies completed in prior years.Figure 9. Distribution of abnormal churn rates in ascending order by per capita costsRegression = Intercept + {abnormal churn rate) x !, where ! denotes the slope.! 10,132,750! 0! 2,000,000! 4,000,000! 6,000,000! 8,000,000! 10,000,000! 12,000,0001 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26Ascending order by the size of data breachTotal Regression! 211! 0! 50! 100! 150! 200! 2501 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26Ascending order by abnormal churn ratesPer Capita Regression
  11. 11. Ponemon Institute© Research Report Page 10Certain industries are more vulnerable to churn. Figure 10 reports the abnormal churn rate ofbenchmarked organizations for the 2012 study. While a small sample size prevents us fromgeneralizing the affect of industry on churn rates, our present industry results are consistent withprior years – wherein financial service organizations experienced relatively high abnormal churnand retail companies and public service entities experienced a relatively low abnormal churn.8Inthis year’s study, public service (government) organizations realize the lowest churn rates.Figure 10. Abnormal churn rates by industry classification of benchmarked companies8Public sector organizations utilize a different churn framework given that customers of government organizations typicallydo not have an alternative choice.1.0%1.8%2.6%5.4%5.4%5.8%5.9%6.0%6.3%7.1%0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% 7.0% 8.0%Public sectorTechnologyRetailConsumerPharmaceuticalHospitalityEnergyIndustrialServicesFinancial
  12. 12. Ponemon Institute© Research Report Page 11Detection and escalation costs decrease slightly. Figure 11 shows the distribution of costsassociated with detection and escalation of the data breach event. Such costs typically includeforensic and investigative activities, assessment and audit services, crisis team management,and communications to executive management and board of directors. As noted, averagedetection and escalation costs decreased from !.75 million in 2011 to !.72 million in 2012.Figure 11. Average detection and escalation costs over four years!000,000 omittedNotification costs increase. Figure 12 reports the distribution of notification costs. Such costsinclude IT activities associated with creation of contact databases, determination of all regulatoryrequirements, engagement of outside experts, postal expenditures, secondary contacts to mail oremail bounce-backs and inbound communication set-up. This year’s average cost of notificationwas !.11. These costs have been relatively stable over the past four years.Figure 12. Average notification costs over four years!000,000 omitted! 0.58 ! 0.58! 0.75! 0.72! 0.00! 0.10! 0.20! 0.30! 0.40! 0.50! 0.60! 0.70! 0.802009 (17) 2010 (21) 2011 (23) 2012 (26)Detection & escalation Average! 0.09! 0.11! 0.10! 0.11! 0.00! 0.02! 0.04! 0.06! 0.08! 0.10! 0.122009 (17) 2010 (21) 2011 (23) 2012 (26)Notification Average
  13. 13. Ponemon Institute© Research Report Page 12Post data breach costs have declined. Figure 13 shows the distribution of costs associatedwith ex-post (after-the-fact) activities. Such costs typically include help desk activities, inboundcommunications, special investigative activities, remediation activities, legal expenditures,product discounts, identity protection services and regulatory interventions. Average ex-postresponse cost decreased from !.91 million in 2011 to !.83 million in this year’s study.Figure 13. Average ex-post response costs over four years!000,000 omittedLost business costs increased sharply. Figure 14 reveals how lost business costs associatedwith data breach incidents have become more significant over the past four years. This costcategory typically includes the abnormal turnover of customers, increased customer acquisitionactivities, reputation losses and diminished goodwill. Lost business costs increased from !.78 in2011 to !1.19 in the present year.Figure 14. Average lost business costs over four years!000,000 omitted! 0.65! 0.82! 0.91! 0.83! 0.00! 0.10! 0.20! 0.30! 0.40! 0.50! 0.60! 0.70! 0.80! 0.90! 1.002009 (17) 2010 (21) 2011 (23) 2012 (26)Ex-post response Average! 0.58! 0.69! 0.78! 1.19! 0.00! 0.20! 0.40! 0.60! 0.80! 1.00! 1.20! 1.402009 (17) 2010 (21) 2011 (23) 2012 (26)Lost business Average
  14. 14. Ponemon Institute© Research Report Page 13The proportion of indirect costs increased. Figure 15 reports the direct and indirect costcomponents of data breach on a per capita basis. In essence, the cost of data breach percompromised record increased by more than !5 – from !122 in 2011 to !127 in 2012.Approximately, !4 of this increase pertains to direct costs. In the present study, indirect costsrepresented 56 percent of total per capita cost.Figure 15. Direct and indirect per capita data breach cost over four years! 30! 39! 52 ! 56! 59! 58! 70! 71! 0! 20! 40! 60! 80! 100! 120! 1402009 (17) 2010 (21) 2011 (23) 2012 (26)Direct cost Indirect cost
  15. 15. Ponemon Institute© Research Report Page 14Preventive measures taken after the breachIn addition to measuring specific cost activities relating to the leakage of personal information, wereport in Table 1 the preventive measures implemented by companies after the data breach. Themost popular measures or steps taken are: training and awareness programs (51 percent), theexpanded use of encryption (40 percent) and the deployment of data loss prevention (DLP)solutions (32 percent). While additional manual procedures and controls are still among the topmeasures at 31 percent, their use has declined 12 percent since last year. Strengthening ofperimeter controls decreased by 5 percent. Other system control practices increased by 5 percentsince the last study.Table 1. Preventive measures andcontrols implemented after the databreach 2009 2010 2011 2012Training and awareness programs 46% 44% 48% 51%Expanded use of encryption 25% 28% 37% 40%Data loss prevention (DLP) solutions 10% 18% 30% 32%Additional manual procedures & controls 53% 51% 43% 31%Security intelligence systems 5% 10% 25% 26%Strengthening of perimeter controls 21% 23% 30% 25%Other system control practices 17% 21% 20% 25%Security certification or audit 20% 19% 22% 23%Endpoint security solutions 8% 16% 23% 20%Identity and access managementsolutions 13% 14% 19% 20%Please note that a company may be implementing more than one preventive measure.
  16. 16. Ponemon Institute© Research Report Page 15Cost changes of data breach categories over timeTable 2 provides the percentage changes for 11 cost categories over four years. As can be seen,most cost categories appear to be relatively stable over time. The two highest cost categoriesare investigation and forensics and lost customer business.Table 2. Percentage data breach costcategories 2009 2010 2011 2012Investigations & forensics 28% 26% 27% 30%Audit and consulting services 13% 9% 10% 9%Outbound contact costs 10% 9% 10% 9%Inbound contact costs 7% 8% 9% 7%Public relations/communications 3% 4% 2% 1%Legal services – defense 1% 3% 3% 4%Legal services – compliance 6% 6% 5% 3%Free or discounted services 2% 1% 1% 2%Identity protection services 0% 1% 0% 1%Lost customer business 23% 25% 26% 27%Customer acquisition cost 7% 7% 7% 7%
  17. 17. Ponemon Institute© Research Report Page 16Part 3. Concluding observations and description about participating companiesFor the first time, companies participating in our annual study report that their data breaches weresmaller in scale and resulted in a lower rate of churn. We conclude that companies’ investment inimproving their data protection practices is paying off. The most profitable investments asevidenced by the lower cost of a data breach are: achieving a strong security posture, appointinga CISO with enterprise-wide responsibility and engaging external consultants.We hope this study helps to understand what the potential costs of a data breach could be basedupon certain characteristics and how best to allocate resources to the prevention, detection andresolution of a data breach. Specifically, the research reveals the severe financial consequencesfrom malicious or criminal acts. These data breaches can prove to be the most costly.In this report, we compare the results of the present study to those from prior years. It isimportant to note that each annual study involves a different sample of companies. In otherwords, we are not tracking the same sample of companies over time. To be consistent, weattempt to recruit and match companies with similar characteristics such as the company’sindustry, headcount, geographic footprint and size of data breach.Figure 16 shows the distribution of benchmark organizations by their primary industryclassification. In this year’s study, 10 industries are represented. Financial services, public sector(government), retail and consumer represent the four largest segments.Figure 16. Distribution of the benchmark sample by industry segment23%19%15%11%8%8%4%4%4%4%Public sectorFinancialRetailConsumerServicesTechnologyEnergyHospitalityIndustrialPharmaceutical
  18. 18. Ponemon Institute© Research Report Page 17Part 4. How we calculate the cost of a data breachOur study addresses core process-related activities that drive a range of expenditures associatedwith an organization’s data breach detection, response, containment and remediation. The fourcost centers are:! Detection or discovery: Activities that enable a company to reasonably detect the breach ofpersonal data either at risk (in storage) or in motion.! Escalation: Activities necessary to report the breach of protected information to appropriatepersonnel within a specified time period.! Notification: Activities that enable the company to notify data subjects with a letter, outboundtelephone call, e-mail or general notice that personal information was lost or stolen.! Ex-post response: Activities to help victims of a breach communicate with the company toask additional questions or obtain recommendations in order to minimize potential harms.Redress activities also include ex-post response such as credit report monitoring or thereissuing of a new account (or credit card).In addition to the above process-related activities, most companies experience opportunity costsassociated with the breach incident, which results from diminished trust or confidence by presentand future customers. Accordingly, our Institute’s research shows that the negative publicityassociated with a data breach incident causes reputation effects that may result in abnormalturnover or churn rates as well as a diminished rate for new customer acquisitions.To extrapolate these opportunity costs, we use a cost estimation method that relies on the“lifetime value” of an average customer as defined for each participating organization.! Turnover of existing customers: The estimated number of customers who will most likelyterminate their relationship as a result of the breach incident. The incremental loss isabnormal turnover attributable to the breach incident. This number is an annual percentage,which is based on estimates provided by management during the benchmark interviewprocess.9! Diminished customer acquisition: The estimated number of target customers who will nothave a relationship with the organization as a consequence of the breach. This number isprovided as an annual percentage.We acknowledge that the loss of non-customer data, such as employee records, may not impactan organization’s churn or turnover.10In these cases, we would expect the business costcategory to be lower when data breaches do not involve customer or consumer data (includingpayment transactional information).9In several instances, turnover is partial, wherein breach victims still continued their relationship with thebreached organization, but the volume of customer activity actually declines. This partial decline isespecially salient in certain industries – such as financial services or public sector entities – wheretermination is costly or economically infeasible.10In this study, we consider citizen, patient and student information as customer data.
  19. 19. Ponemon Institute© Research Report Page 18All participating organizations experienced one or more data breach incidents sometime over thepast year. Our benchmark instrument captured descriptive information from IT, compliance andinformation security practitioners about the full cost impact of a breach involving the loss or theftof customer or consumer information. It also required these practitioners to estimate opportunitycosts associated with program activities.Estimated data breach cost components were captured on a rating form. In most cases, theresearcher conducted follow-up interviews to obtain additional facts, including estimatedabnormal churn rates that resulted from the company’s most recent breach event involving 1,000or more compromised records.11Data collection methods did not include actual accounting information, but instead relied upon anumerical estimation based upon the knowledge and experience of each participant. Within eachcategory, cost estimation was a two-stage process. First, the benchmark instrument requiredindividuals to rate direct cost estimates for each cost category by marking a range variabledefined in the following number line format.How to use the number line: The number line provided under each data breach cost category is one way toobtain your best estimate for the sum of cash outlays, labor and overhead incurred. Please mark only onepoint somewhere between the lower and upper limits set above. You can reset the lower and upper limitsof the number line at any time during the interview process.Post your estimate of direct costs here for [presented cost category]LL ______________________________________|___________________________________ ULThe numerical value obtained from the number line rather than a point estimate for eachpresented cost category preserved confidentiality and ensured a higher response rate. Thebenchmark instrument also required practitioners to provide a second estimate for indirect andopportunity costs, separately.The scope of data breach cost items contained within our benchmark instrument is limited toknown cost categories that are applied to a broad set of business operations that handle personalinformation. We believe a study focused on business process – and not data protection or privacycompliance activities – yields a better quality of results.11Our sampling criteria only included companies experiencing a data breach between 1,000 and 100,000lost or stolen records sometime during the past 12 months. We excluded catastrophic data breach incidentsto avoid skewing overall sample findings.
  20. 20. Ponemon Institute© Research Report Page 19Figure 17 illustrates the activity-based costing schema used in our benchmark study. The costcenters we examine sequentially are: incident discovery, escalation, notification, ex-postresponse and lost business.Figure 17. Schema of the data breach processWithin each cost center, the research instrument required subjects to estimate a cost range tocapture estimates of direct cost, indirect cost and opportunity cost, defined as follows:! Direct cost – the direct expense outlay to accomplish a given activity.! Indirect cost – the amount of time, effort and other organizational resources spent, but not asa direct cash outlay.! Opportunity cost – the cost resulting from lost business opportunities as a consequence ofnegative reputation effects after the breach has been reported to victims (and publiclyrevealed to the media).To maintain complete confidentiality, the benchmark instrument did not capture any company-specific information. Subject materials contained no tracking codes or other methods that couldlink responses to participating companies.To keep the benchmarking process to a manageable size, we carefully limited items to only thosecost activity centers that we considered crucial to data breach cost measurement. Based upondiscussions with learned experts, the final set of items included a fixed set of cost activities. Uponcollection of the benchmark information, each instrument was re-examined carefully forconsistency and completeness.
  21. 21. Ponemon Institute© Research Report Page 20LimitationsOur study utilizes a confidential and proprietary benchmark method that has been successfullydeployed in earlier research. However, there are inherent limitations with this benchmarkresearch that need to be carefully considered before drawing conclusions from findings.! Non-statistical results: Our study draws upon a representative, non-statistical sample ofFrench-based entities experiencing a breach involving the loss or theft of customer orconsumer records during the past 12 months. Statistical inferences, margins of error andconfidence intervals cannot be applied to these data given that our sampling methods are notscientific.! Non-response: The current findings are based on a small representative sample ofbenchmarks. Twenty-six companies completed the benchmark process. Non-response biaswas not tested so it is always possible companies that did not participate are substantiallydifferent in terms of underlying data breach cost.! Sampling-frame bias: Because our sampling frame is judgmental, the quality of results isinfluenced by the degree to which the frame is representative of the population of companiesbeing studied. It is our belief that the current sampling frame is biased toward companieswith more mature privacy or information security programs.! Company-specific information: The benchmark information is sensitive and confidential.Thus, the current instrument does not capture company-identifying information. It also allowsindividuals to use categorical response variables to disclose demographic information aboutthe company and industry category.! Unmeasured factors: To keep the interview script concise and focused, we decided to omitother important variables from our analyses such as leading trends and organizationalcharacteristics. The extent to which omitted variables might explain benchmark resultscannot be determined.! Extrapolated cost results. The quality of benchmark research is based on the integrity ofconfidential responses provided by respondents in participating companies. While certainchecks and balances can be incorporated into the benchmark process, there is always thepossibility that respondents did not provide accurate or truthful responses. In addition, theuse of cost extrapolation methods rather than actual cost data may inadvertently introducebias and inaccuracies.
  22. 22. Ponemon Institute© Research Report Page 21Appendix 1: Cost for 26 Data Breach Case StudiesCasesSize ofbreachDetection &escalation* Notification*Ex-postresponse*Lostbusiness* Total*1 18,457 249,087 54,219 926,652 658,415 1,888,3732 10,989 249,426 18,247 687,297 965,959 1,920,9293 20,162 452,727 3,974 362,006 838,869 1,657,5764 44,355 1,171,944 106,298 1,232,099 1,114,696 3,625,0375 5,391 172,103 87,644 26,266 29,300 315,3136 4,470 290,850 12,015 227,058 234,393 764,3167 24,345 784,486 163,252 910,888 1,882,971 3,741,5978 32,323 1,139,206 239,816 1,204,208 2,615,553 5,198,7839 33,027 889,155 248,825 1,218,023 2,379,030 4,735,03310 19,212 590,194 4,470 436,268 50,331 1,081,26311 25,796 505,933 248,620 1,952,694 2,733,043 5,440,29012 19,473 710,103 6,167 63,011 798,754 1,578,03513 14,038 199,516 144,632 674,741 432,927 1,451,81614 72,186 2,499,736 167,974 2,867,572 4,597,468 10,132,75015 38,321 1,013,380 154,014 1,993,836 1,189,551 4,350,78116 41,810 2,310,692 343,178 1,911,453 2,597,133 7,162,45617 4,620 173,460 26,990 128,973 134,043 463,46618 17,461 343,382 201,728 379,091 11,662 935,86319 12,877 521,608 112,557 502,124 1,149,166 2,285,45520 55,394 2,029,906 220,301 2,257,417 3,563,018 8,070,64221 14,167 438,891 54,095 247,004 14,157 754,14722 16,367 834,487 96,530 313,661 1,261,045 2,505,72323 16,104 379,241 4,542 529,379 629,266 1,542,42824 14,444 475,797 98,510 354,727 943,478 1,872,51225 2,381 51,246 48,642 18,637 120,906 239,43126 5,851 356,576 12,556 202,511 77,132 648,775*Measured in Euros (!)
  23. 23. Ponemon Institute© Research Report Page 22If you have questions or comments about this research report or you would like to obtainadditional copies of the document (including permission to quote or reuse this report), pleasecontact by letter, phone call or email:Ponemon Institute LLCAttn: Research Department2308 US 31 NorthTraverse City, Michigan 49686 USA1.800.887.3118research@ponemon.orgPonemon Institute LLCAdvancing Responsible Information ManagementPonemon Institute is dedicated to independent research and education that advances responsibleinformation and privacy management practices within business and government. Our mission isto conduct high quality, empirical studies on critical issues affecting the management and securityof sensitive information about people and organizations.As a member of the Council of American Survey Research Organizations (CASRO), weuphold strict data confidentiality, privacy and ethical research standards. We do not collect anypersonally identifiable information from individuals (or company identifiable information in ourbusiness research). Furthermore, we have strict quality standards to ensure that subjects are notasked extraneous, irrelevant or improper questions.