ORDER
THOMAS B. SMITH, Magistrate Judge.
Following a hearing, the Court enters this Order on Plaintiff's Motion to Compel Discovery Responses from Defendants (Doc. 80). Plaintiff Deborah Clark alleges that Defendants FDS Bank ("FDS") and Department Stores National Bank ("DSNB") violated the Telephone Consumer Protection Act ("TCPA"), 47 U.S.C. § 22 et seq. by using an automated telephone dialing system to call her cell phone approximately 50 times without her consent, in an attempt to collect a debt owed on her daughter's Macy's credit card1 (Doc. 53). Plaintiff brings this action on her own behalf and on behalf of the following proposed classes:
The TCPA Class:
(1) All persons in the United States (2) to whose cellular telephone number (3) FDS placed a non-emergency telephone call on behalf of DSNB (4) using substantially the same system(s) that were used to telephone Plaintiff (5) within 4 years of the complaint and (6) where Defendants did not obtain the cellular telephone number from the consumer who was called.
The DSNB Sub-Class:
(1) All persons in the United States (2) to whose cellular telephone number (3) DSNB placed a non-emergency telephone call (4) using substantially the same system(s) that were used to telephone Plaintiff (5) within 4 years of the complaint and (6) where Defendants did not obtain the cellular telephone number from the consumer who was called.
The Revocation Sub-Class:
(1) All persons in the United States (2) to whose cellular telephone number (3) DSNB and/or FDS placed a non-emergency telephone call (4) using substantially the same system(s) that were used to telephone Plaintiff (5) within 4 years of the complaint and (6) after that person had instructed Defendants to cease calls to that number.
The Skip-Trace Sub-Class:
(1) All persons in the United States (2) to whose cellular telephone number (3) DSNB and/or FDS placed a non-emergency telephone call (4) wherein the cellular telephone number was obtained through skip-tracing (5) using substantially the same system(s) that were used to telephone Plaintiff (6) within 4 years of the complaint.
(Id., ¶ 57).
Plaintiff wants to know the following information about collection calls made by Defendants during a four year period: (1) the phone numbers called; (2) the dialing system used to make each call; (3) the date on which each call was made; (4) the number of times each phone number was called; (5) where Defendants got the phone numbers they called; (6) for each call, whether it produced a live answer, prerecorded message, or report of a wrong number; (7) whether the recipient of the call told Defendants not to call the phone number again; (8) the account notes concerning each call; and (9) financial information to show when payments for the accounts associated with each phone number were made (Doc. 98 at 5-9).
Defendants make approximately 137,000 telephone calls per month to account holders (Doc. 98 at 13). They represent that the account notes associated with all of the calls made during the last four years total in the tens of millions (Doc. 89 at 12; Doc. 98 at 27). Defendants also advise that the account notes do not include all of their databases and documents that might have to be searched to locate the information Plaintiff wants (Doc. 89 at 12). Defendants maintain the ledgers showing what is owed, how much is paid and when in a different system which also contains account notes (Doc. 98 at 18). Defendants can download from the dialing system the total number of calls made, and they can produce a list of the phone numbers called, the dates and times the calls were placed and if the call was answered, the duration of the call (Doc. 98 at 15-16, 18, 20). But, according to Defendants, to obtain all of the information requested they must pull up each individual account and manually review it (Id.). They are adamant that the account notes are only searchable manually (Id., at 28). During a FED. R. CIV. P. 30(b)(6) deposition, a Director of Collections for Macy's Credit Operations, Inc., a wholly-owned subsidiary of FDS testified:
So the system is not searchable. I cannot — as a collector, when I worked accounts, I am not able to — I have to scroll through every note. I have to review every single note as far back as I wanted to go and needed to go. There's no way to pull specific aspects of the note.
(Doc. 122-1, Pg. 30, Lines 5-10). A legal assistant in the Macy's Law Department was asked to retrieve account information for 28 randomly selected accounts that were in collections in July 2015 (Doc. 114, ¶¶ 1, 3). Her declaration states:
5. After inputting the account information, the credit information database displays the relevant account. I then selected the fields I wished the system to retrieve, here the account and collection notes. Once retrieved, I printed the account and collection notes (the print to PDF and "screen print" functionality is disabled in the system), retrieved them, and scanned them into a PDF document, before saving the file electronically. I repeated this process for each of the remaining twenty-seven (27) accounts.
6. In performing this function I kept track of the time I spent accessing accounts, printing the account notes and collection notes, scanning them into a PDF document, and then saving the electronically.
7. In total, I spent approximately 50 minutes locating accounts, printing account and collection notes for the 28 accounts, an additional 14 minutes scanning in the account notes and collection notes, and 13 minutes saving the various PDF files, for a total of approximately 76 minutes.
(Doc. 114, ¶¶ 5-7). Many of the accounts this witness retrieved also have archived collection notes that were not readily available and required additional time to obtain (Id., ¶ 10).
If the information is produced then Defendants say they have obligations to their customers under various banking regulations including the Gramm-Leach-Bliley Act, 15 U.S.C. § 6801 that will require the redaction of customers' nonpublic, personal information (Doc. 103 at 2). It took an administrative assistant in the Macy's Law Department 135 minutes to review and redact personal identifying information from the account notes of 17 people (Doc. 104, ¶¶ 1,3). A paralegal in the same department needed 120 minutes to review and redact personal identifying information from the account notes of 11 individuals (Doc. 104-2, ¶¶ 1, 3). After considering the magnitude of Plaintiff's requests, Defendants estimate the cost to review and redact the information Plaintiff seeks will be between $1,533,169 and $1,864,665 (Doc. 103 at 3).
If these hurdles can be overcome, then Defendants argue that review of the account notes will require them to "make the highly factually intensive determination (and, as applied, legal conclusion) as to whether or not the call recipient instructed calls to stop or provided consent respectively." (Doc. 89 at 6). Defendants also contend that Plaintiff is attempting to create a "fail-safe" scenario that "`front-ends a merit determination on the [defendants'] liability'" by requiring the identification of "`all persons . . . who were called after they requested that defendant stop calling them.'" (Id.) (quoting Tillman v. Ally Financial Inc., Case No. 2:16-cv-313-FtM-99CM, 2016 WL 6996113, at *5 (M.D. Fla. Nov. 30, 2016) (quoting Kamar v. RadioShack Corp., 375 Fed. Appx. 734, 736 (9th Cir. 2010) ("The fail-safe appellation is simply a way of labeling the obvious problems that exist when the class itself is defined in a way that precludes membership unless the liability of the defendant is established. When the class is so defined, once it is determined that a person, who is a possible class member, cannot prevail against the defendant, that member drops out of the class. That is palpably unfair to the defendant, and is also unmanageable-for example, to whom should the class notice be sent?")).
Plaintiff does not accept Defendants' protestations concerning the burden her requested discovery will impose on them. She has filed the Declaration of Jeffrey A. Hansen in support of her contention that the information she seeks is readily accessible (Doc. 122-2). Mr. Hansen is the principal of Hansen Legal Technologies, Inc., which, among other things, investigates and analyzes electronic data (Id., ¶ 4). He claims "extensive experience dealing with data warehousing, including data warehousing related to telemarketing and autodialers in general." (Id., ¶ 5). Mr. Hansen says he is familiar "with the procedures involved in such practices, and I have personally engaged in data warehousing regarding the compilation of certain lists, including demographic and target audience lists for telemarketing, and have personally repaired defective lists to eliminate improperly formatted and corrupted data." (Id.). Mr. Hansen also claims knowledge of the autodialer used by Defendants, the design of the databases within the dialer, and the information that should be contained within those databases (Id., ¶¶ 11-12). He states that the databases in the autodialer maintain pertinent "records in a very easy to use, and easy to export format." (Id., ¶ 12). Mr. Hansen "estimate[s] it would take 10-15 minutes to export the data from the . . . dialer." (Id., ¶ 13). He also estimates that it "would take less than an hour to export" the data Plaintiff wants from all of the collection notes (Id., ¶ 14).
Attached to Mr. Hansen's declaration is a copy of the Declaration of Robert Biggerstaff filed in a different case (Id., at 88). According to Mr. Hansen "[t]hat matter was identical to this matter as far as the data analysis." (Id., ¶ 22). Mr. Hansen also declares that "all of Robert Biggerstaff's declaration is relevant to this matter." (Id.) (emphasis in original). In his declaration Mr. Biggerstaff states that he is:
a certified computer forensic examiner and a member of the International Society of Forensic Computer Examiners ("ISFCE"). I have over 30 years of experience in forensic computer examinations and data recovery. I have been retained as a expert in computer forensics and data analysis in over 300 court cases and have performed hundreds of data recovery and ESI preservation assignments across many diverse platforms and filesystems.
(Id., at 89). Mr. Biggerstaff represents that he has "direct experience in the role that computer records play in claims under the Telephone Consumer Protection Act ("TCPA"), specifically with regard to analyzing those records for the purposes of identifying persons to whom telephone calls were made, the circumstances of those calls, and for purposes of class certification and class notice." (Id.). According to Mr. Biggerstaff:
12. I am aware that some people (even IT professionals directly involved with the data) using business databases similar to those used by Ally believe that their records can not be accessed or searched to produce certain information, particularly when the data sought is recorded in "free-form" text fields. However, I have been able to extract information and generate reports from such computer records that the users of the system were genuinely unaware could be obtained, or believed could only be obtained with great difficulty. Similarly, information that was believed to require extensive programming or machine-time to obtain, was in fact obtainable by me in less than an hour's work.
32. The aggregate size of the data is really only a minor issue. The amount of work to write a script to export the data is largely the same whether it is exporting 10,000 records or 1 billion records. The latter will just take longer to run once it starts. The same is true for my analysis—I do essentially the same amount of work whether I am analyzing 10,000 records or a billion records. This type of analysis is simple to do and I have done it before in many cases and with larger databases.
33. One reason is that my forensic analysis system is optimized and specially tailored for forensic analysis, and not intended for day-to-day business operation. Suppose a dialer database has 100 million records to search for text mentioning the words "wrong number." Such a search may take several hours on a particular unindexed system that is designed for day-to-day business use and not optimized for such a free-text search. But that same search can take just seconds on an optimized system like I operate for forensic analysis, once the data is loaded onto that system.
34. The same business system that may have 100 million records, likely contains many columns of data for each record. My task (to identify records where the notes indicate a DNC request or wrong number occurred) needs only a few columns of data (primarily account number, date, phone number, and notes for my analysis.) Exporting any table (relational or not) to CSV or other portable format is a simple task supported by all conventional database platforms. Ally can simply export those columns of data to one or more external CSV files, and send them to me to analyze and conduct the searches. I have received and analyzed data using this exact procedure in many cases.
36. I can easily load hundreds of millions text records into a work index or forensic tool and be able to do very sophisticated searches of the entire corpus of those records practically instantly. This allows me to conduct a robust analysis of the data using proximity operators, wildcards, word stemming, and other techniques.
37. By examining the entire body of those records at once, I will be able to efficiently develop an objective algorithm to identify records containing DNC requests and wrong number calls. That objective algorithm will then be able to be examined by anyone for its accuracy and error rate (if any) against the data.
(Id., at 91, 97-98) (emphasis in original) (footnote omitted).
Plaintiff's evidence is sufficient to persuade the Court that Defendants' business records may not be as difficult to search as they contend. Accordingly, the motion to compel is GRANTED as follows.
First, the parties should be guided by the following practice pointers:
The successful use of any automated search method or technology will be enhanced by a well-thought-out process with substantial human input on the front end.
Parties and their counsel should make a good faith attempt to cooperate when determining the use of particular search and information retrieval methods, tools, and protocols (including keywords, concepts, computer—or technology—assisted review and other types of search parameters and quality control measures.
The Sedona Conference Journal Volume 15, Fall 2014, The Sedona Conference Best Practices Commentary on the Use of Search & Information Retrieval Methods in E-Discovery pages 224-225.
With these pointers in mind, counsel and the parties' experts shall hold a FED. R. CIV. P. 26(f) conference within 21 days from the rendition of this Order. At the conference they will discuss how Defendants' business records will be made available for search and retrieval by Mr. Hansen. They will also discuss how information not included in items (1)-(9) above, or that is confidential, is removed and redacted from the production to be made to Plaintiff. The Court expects Plaintiff to identify the specific fields of information she needs and agree to the use of anonymizers to conceal account holders' identities and other personal information. The Court also expects Plaintiff to agree to reasonable precautions to protect the integrity and confidentiality of Defendants' systems and procedures. In like manner, the Court expects Defendants to agree on a reasonable methodology that gives Mr. Hansen the access he needs to attempt to search and retrieve relevant information. This may mean that Mr. Hansen is given direct access to Defendants' systems, that he is provided with mirror images, or that some other reasonable methodology is employed. All of these activities will be conducted subject to an appropriate confidentiality agreement signed by all involved on the Plaintiff's side. If the parties cannot resolve these matters between themselves within the next 35 days, then on motion filed the Court will hold an evidentiary hearing where it will hear from the experts and then dictate the process and procedure to the parties.
Defendants' concerns about the subjective interpretation of the account note entries and the creation of "fail-safe" classes are issues to be dealt with later in the case. For now, discovery must get underway so that the case can move forward.
All other relief requested in the motion to compel is DENIED without prejudice.
DONE and ORDERED.
