Revitch v. Citibank, N.A.

Filing 150

ORDER RE 84 MOTION FOR CLASS CERTIFICATION by Judge Alsup. (whalc1S, COURT STAFF) (Filed on 4/28/2019)

1
2
3
4
5
6

IN THE UNITED STATES DISTRICT COURT

7
FOR THE NORTHERN DISTRICT OF CALIFORNIA
8
9

11
For the Northern District of California

United States District Court

10

JEREMIAH REVITCH, on behalf of
himself and all others similarly situated,

12

Plaintiff,

13
14
15

No. C 17-06907 WHA
v.

ORDER RE MOTION FOR
CLASS CERTIFICATION

CITIBANK, N.A.,
Defendant.
/

16
17
18
19
20

INTRODUCTION
In this putative class action for alleged violations of the Telephone Consumer Protection
Act, plaintiff moves for class certification. For the reasons herein, the motion is DENIED.
STATEMENT

21

Plaintiff Jeremiah Revitch filed this putative class action in December 2017, alleging

22

that defendant Citibank, N.A. called him multiple times on his cell phone using an autodialer

23

despite plaintiff not being a Citibank customer. Plaintiff continued to receive these calls even

24

after he told Citibank that it had reached a wrong number. Based on these allegations, plaintiff

25

asserted two claims for negligent and willful violations of the Telephone Consumer Protection

26

Act. Plaintiff now moves to certify one class pursuant to FRCP 23(b)(2) and 23(b)(3). On

27

reply, plaintiff amended his proposed class definition to the following:

28

Jeremiah Revitch and all persons in the United States (1) whose
cellular telephone is identified in Defendant’s Contact Utilities
Database; (2) who between March 17, 2014, through August 21,
2018; (3) were called on their cellular telephone by Defendant or

1

its agent/s using its Aspect Unified dialer; and (4) where such
person was not listed in Defendant’s records as the intended
recipient of the calls.

2
3

Following a February 2019 hearing on plaintiff’s motion for class certification, an order

4

allowed supplemental briefing to permit Citibank an opportunity to address information raised

5

by plaintiff for the first time on reply. The parties filed supplemental briefs and evidence and

6

the undersigned held a second hearing on the motion in April 2019 (Dkt. Nos. 135, 137). This

7

order accordingly follows two rounds of briefing and oral argument.1

8

ANALYSIS

9

Federal Rule of Civil Procedure 23(a) provides that a member of a class may sue as a
representative party on behalf of all members where: (1) the class is so numerous that joinder

11
For the Northern District of California

United States District Court

10

of all members is impracticable; (2) there are questions of law or fact common to the class; (3)

12

the claims or defenses of the representative parties are typical of the claims or defenses of the

13

class; and (4) the representative parties will fairly and adequately protect the interests of the

14

class. FRCP 23(b) sets forth three conditions under which, if the prerequisites of FRCP 23(a)

15

are satisfied, a class action may be maintained. Class certification is only appropriate if a

16

plaintiff meets all the prerequisites of FRCP 23(a) and at least one condition of FRCP 23(b).

17

Abdullah v. U.S. Sec. Assocs., Inc., 731 F.3d 952, 956–57 (9th Cir. 2013). In contrast to the less

18

stringent standard of FRCP 23(a)(2), class certification under FRCP 23(b)(3) is proper only

19

when common questions present a significant portion of the case and can be resolved for all

20

class members in a single adjudication. Hanlon v. Chrysler Corp., 150 F.3d 1011, 1022 (9th

21

Cir. 1998).
The TCPA prohibits making unsolicited calls to cellular phones using an automatic

22
23

dialing system. 47 U.S.C. § 227(b)(1)(A)(iii). A call is not unsolicited, however, where the

24

“called party” gave the sender “prior express consent.” Id. at § 227(b)(1)(A). Citibank issues

25

and services credit cards. In the course of business, account holders provide their phone

26

numbers to Citibank in various ways, including through applications, online, or during calls

27
28

1

Citibank’s administrative motion for leave to supplement the record with the entirety of Daley’s
deposition testimony is unopposed and is accordingly GRANTED.

2

1

with Citibank employees (Meeks Decl. ¶ 5; Meeks Dep. 31:3–33:9). Plaintiff concedes that

2

Citibank had prior express consent to call its customers about their accounts (Dkt. No. 137 at 1).

3

Accordingly, plaintiff attempts to define the proposed class such that it includes only

4

individuals who received calls from Citibank about someone else’s account. Despite these

5

efforts, this order concludes that individualized issues of consent will predominate at trial,

6

making class certification impractical in this case.

7

For each customer account, Citibank maintains various records, including customer

8

service notes, collection notes, and customer contact information. When an account goes

9

delinquent, Citibank calls the account holder at the number listed in its records. To make these
calls, Citibank uses a system from its vendor Aspect Software, Inc., through which call

11
For the Northern District of California

United States District Court

10

operators flag phone numbers with various codes related to the customer’s consent to call. For

12

example, call operators can flag a phone number as a “bad number” or “not valid” when they

13

believe they have reached a number that does not belong to the customer they attempted to

14

reach. Citibank maintains this “phone indicator data” in a database along with the name of the

15

intended recipient and associated phone number. Starting in November 2017, Citibank began to

16

track the historical data associated with these consent codes, thereby showing any prior codes

17

associated with an account at the time a call was made (Roe Decl. ¶¶ 4–5; Mullahey Decl. ¶¶

18

7–14).

19

As plaintiff recognizes, a phone number can be flagged as “wrong” in Citibank’s system

20

even when it is, in fact, the customer’s correct number. For example, an evasive customer may

21

reply with “wrong number” when he answers a call regarding his delinquent account. Plaintiff

22

claims to account for this issue by conducting a “reverse lookup” using TransUnion TLOxp

23

data. Specifically, plaintiff proposes the following methodology for identifying class members

24

who received “wrong number” calls from Citibank on their cellular phones. First, identify from

25

Citibank’s records any phone number marked with a letter-code indicating a “wrong number.”

26

Second, filter out any number that was once marked as a “wrong number” but has since been

27

given a flag indicating that the call recipient later said something to the contrary. Third, use

28

information from TransUnion to exclude any calls made to landline numbers. Fourth, use

3

1

TransUnion’s historical database lookup service to identify the customary user of a phone

2

number during the time period that Citibank’s call was made. Where the name of the user

3

identified from this “reverse lookup service” does not match the name of the account holder

4

listed in Citibank’s records, plaintiff argues you have confirmed a true “wrong number” and

5

have identified a member of the class. Based on a sample set of 20,000 phone numbers,

6

plaintiff’s expert originally identified 176 phone numbers that fell within the definition of the

7

proposed class using this methodology. In applying the revised methodology he developed for

8

plaintiff’s reply brief, this number dropped to 133 phone numbers (Snyder Decl. ¶¶ 68–79; Weir

9

Decl. ¶ 25; Weir Reply Decl. ¶ 64).
Citibank identifies several fatal flaws in plaintiff’s methodology. As an initial matter, of

11
For the Northern District of California

United States District Court

10

the 133 phone numbers plaintiff identified as receiving a “wrong numbers” call, a review of the

12

underlying account data associated with those phone numbers shows that many of them were

13

associated with the Citibank customer that the bank was attempting to reach. For example, one

14

such phone number was provided to Citibank in connection with a father’s account. Over a

15

year later, his son provided the same phone number to Citibank in connection with a different

16

account. When Citibank later called that phone number to reach the father, his son answered

17

and explained that it was no longer a good number at which to reach his father. Citibank

18

marked the number as a “bad number,” indicating that the number had been answered by a third

19

party. A few months later, however, Citibank made contact with the father at that same phone

20

number, indicating that the prior call was not, in fact, made to a “wrong number” and would not

21

have been properly included within the proposed class (Daley Decl. ¶¶ 125–30).

22

Notably, a single phone number can be associated with multiple accounts owned by

23

different people. Yet, if a number is marked “wrong” on any of these accounts, plaintiff’s

24

methodology considers every future call to that number “wrong” regardless of which account

25

was called and despite the fact that a phone number may be “wrong” for one account or person

26

but valid for a different account or person. As a result, in reviewing the account-specific data

27

for the phone numbers plaintiff identified through his methodology, Citibank’s expert

28

determined that many of these phone numbers received calls only in connection with a different

4

1

account than the account where the number was marked as a “wrong number” (Daley Decl. ¶

2

152).2
Turning to another problem with plaintiff’s methodology, TransUnion TLOxp reports

3
4

the names of people associated with phone numbers as well as the dates pertaining to those

5

associations. This data includes both a “first seen” date and a “last seen” date. The “first seen”

6

date identifies the earliest date an individual’s name is associated with a phone number. The

7

“last seen” date represents the last time an individual’s name was associated with a number.

8

Together with the “first seen” date, this forms a bracket of time during which there is some kind

9

of documentation indicating an association between the name and number, with a longer
bracket of time indicating a likelier name/number association. In applying plaintiff’s

11
For the Northern District of California

United States District Court

10

methodology, however, his expert did not examine or evaluate the “last seen” date at all.

12

Taking into account this additional data, Citibank’s expert determined that 120 of the original

13

176 phone numbers identified as class members through plaintiff’s methodology were actually

14

documented by TransUnion as having been associated with the Citibank customer named on the

15

account at the time of the call. Plaintiff responds that his methodology need not take into

16

account the “last seen” date because his expert believes it often contains outdated information

17

and that the most recent “first seen” date is a more “conservative” data point, but this ignores

18

that plaintiff’s other expert, Randall Snyder, explained that where the “first seen” and “last

19

seen” dates for different names overlap, the only options are to exclude that phone number from

20

the class “or do further analysis to actually confirm that it should or should not be in the class”

21

(Daley Report ¶¶ 80–90; Snyder Dep. 202:21–203:16; Weir Reply Decl. ¶¶ 31–34).

22

An additional problem with plaintiff’s methodology is that he seeks to represent

23

individuals who received wrong-number calls dating back to March 2014, yet Citibank did not

24

begin to retain historical data concerning changes to wrong-number designations until

25

November 2017. Prior to that date, this information could only be found in account notes, call

26
27
28

2

Although plaintiff complains that he could not have incorporated Citibank’s account numbers into his
methodology because Citibank did not produce the unredacted data until January 11, 2019, plaintiff clearly had
no qualms about fixing other aspects of his methodology in his February 2019 reply brief, such as modifying it
to address Citibank’s criticisms that the name matching step failed to exclude businesses and misidentified as
class members those individuals who may have changed their last name.

5

1

logs, call recordings and similar account-level information. Plaintiff fails to deal with the fact

2

that he seeks to represent a class of individuals who received phone calls dating back to March

3

2014, yet the records upon which he relies in applying his methodology do not predate

4

November 2017. And, contrary to plaintiff’s complaint that Citibank “systematically deleted

5

consent information” prior to this lawsuit, this information remains available. Unfortunately for

6

plaintiff, the format of this data is simply unamenable to an automated review.

7

In light of this problem, plaintiff is forced to propose a class definition that expressly

8

includes himself. Plaintiff’s proposed class now includes “Jeremiah Revitch and all persons in

9

the United States” identified through his methodology. The reason for this oddity is that
plaintiff’s methodology excludes him from the class that he purports to represent. The parties

11
For the Northern District of California

United States District Court

10

seem to agree that plaintiff received a wrong number call. But the only way to prove this at trial

12

would be through individualized account data and testimony. To be sure, this evidence would

13

show that plaintiff’s phone number was mistakenly provided to Citibank by its customer.

14

Citibank then called plaintiff’s number when the customer’s account went delinquent. During

15

the fifth such call in July 2017, plaintiff informed Citibank that the number did not belong to its

16

customer and Citibank’s agent marked the number as “wrong.” Following this conversation,

17

however, Citibank reached the customer on another phone number. During that call, the

18

customer promised to pay his account and the Citibank agent changed the consent flag from

19

“wrong number” to “verbal do not call” as a result. Accordingly, even if the data upon which

20

plaintiff’s methodology relies documented this pre-November 2017 change in consent codes

21

(which it does not), the code change from “wrong” to “no consent” would nevertheless exclude

22

plaintiff as a class member (Exelrod Dep. 13:1–14:19; Mullahey Decl. ¶¶ 15–23; Roe Decl. ¶

23

8).

24

District courts have split on the propriety of class certification in similar “wrong

25

number” TCPA actions. Plaintiff heavily relies on a line of decisions that includes West v.

26

California Services Bureau, Inc., 323 F.R.D. 295 (N.D. Cal. 2017), where Judge Yvonne

27

Gonzalez Rogers granted certification of an similarly-defined “wrong number” class in a TCPA

28

action brought again a debt-collection company. There, as here, the defendant denied that

6

1

phone numbers denoted as “wrong numbers” in the defendant’s system actually reflected calls

2

to wrong numbers. Judge Gonzalez Rogers found persuasive, however, that plaintiff could use

3

a reverse number lookup service to find discrepancies between the service’s and the defendant’s

4

records and thereby resolve consent on a class-wide basis. Id. at 300–02.

5

Citibank, in turn, points to a line of cases that includes Tomeo v. CitiGroup, Inc., No.

6

13-cv-4046, 2018 WL 4627386 (N.D. Ill. Sept. 27, 2018) (Judge Sara Ellis), where the district

7

court concluded that FRCP 23(b)(3)’s predominance requirement had not been met where the

8

defendant put forth evidence establishing that a significant percentage of the putative class

9

consented to receiving calls. Specifically, the defendant’s experts demonstrated that the
individuals coded as “wrong numbers” did not necessarily constitute individuals who did not

11
For the Northern District of California

United States District Court

10

consent to calls. Id. at *9. Plaintiff argues that Tomeo is distinguishable because, there, the

12

plaintiff relied only on the defendant’s records to identify wrong number calls without using a

13

“reverse lookup service” as is proposed in the instant case. Relying on Judge Gonzalez

14

Rogers’s decision in West, plaintiff further argues that this difference is determinative. This

15

order disagrees.

16

Unlike the defendant in West, and as in Tomeo, Citibank has put forward an evidentiary

17

basis from which to conclude that adjudicating whether or not members of the class consented

18

to its calls lacks a common method of proof. The undersigned judge remains convinced that the

19

consent issue will devolve into individualized inquiries which would overwhelm the trial. For

20

this same reason, our court of appeals’ recent decision in True Health Chiropractic, Inc. v.

21

McKesson Corporation, 896 F.3d 923 (9th Cir. 2018), does not aid plaintiff. There, the

22

appellate court recognized that “[a] defendant can produce evidence of a

23

predominance-defeating consent defense in a variety of ways.” Id. at 931–32. Accordingly,

24

and although it refused to consider “consent defenses that [the defendant] might advance or for

25

which it has presented no evidence,” True Health concluded that the defendant had provided

26

evidence in the district court that its consent defenses as to certain claims “would be based on

27

individual communications and personal relationships between [the defendant’s] representatives

28

7

1

and their customers” and that “[t]he variation in such communications and relationships”

2

warranted denying class certification as to those claims. Id. at 932. So too here.

3

Plaintiff’s remaining arguments to the contrary lack merit. Plaintiff incorrectly portrays

identifying the individuals who fall within plaintiff’s proposed class. Rather, the problem is

6

that adjudicating the claims of those who do fall within plaintiff’s proposed class would devolve

7

into the tedious resolution of individualized issues based on individualized evidence. Plaintiff

8

also argues that Citibank could not use its account-level documents to show consent at trial

9

because the documents are inadmissible double hearsay. This order disagrees. To the extent

10

the records contain statements by putative class members indicating consent, those statements

11
For the Northern District of California

consent as an issue of ascertainability rather than predominance. The problem here is not

5

United States District Court

4

are non-hearsay admissions of party opponents. FRE 801(d)(2). As to the account records

12

themselves, plaintiff argues that the records “indicate lack of trustworthiness” such that the

13

records cannot fall with FRE 803(6)’s exception for business records. This order again

14

disagrees. While some portion of Citibank’s records run the risk of containing misinformation

15

due to “human error,” nothing indicates that Citibank’s records are so unreliable that they

16

should be deemed inadmissible in their entirety.3

17

Because plaintiff has failed to meet his burden of demonstrating that certification of a

18

FRCP 23(b)(3) class would be appropriate, the motion for class certification is DENIED. And,

19

because where, as here, “the monetary relief claims are key to this action and cannot reasonably

20

be called incidental,” class certification under FRCP 23(b)(2) is not appropriate. Lozano v. AT

21

& T Wireless Servs., Inc., 504 F.3d 718, 729 (9th Cir. 2007). Plaintiff’s alternative request for

22

class certification under FRCP 23(b)(2) is accordingly DENIED.

23
24
25
26
27
28

3
This order need not (and does not) reach Citibank’s argument that its arbitration agreement with
account holders also precludes class certification. This order also does not reach Citibank’s argument that
plaintiff has failed to demonstrate that FRCP 23(a)’s requirements are met.

8

1
2
3

CONCLUSION
For the foregoing reasons, plaintiff’s motion for class certification is DENIED. This case
will proceed to trial as scheduled as an individual case.

4
5

IT IS SO ORDERED.

6
7

Dated: April 28, 2019.

WILLIAM ALSUP
UNITED STATES DISTRICT JUDGE

8
9

11
For the Northern District of California

United States District Court

10

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

9