PLATFORM

Field

[1] The present disclosure relates generally to methods and apparatus for operating an email demand-side platform.

Background

[2] Proprietors of email address lists may wish to better exploit these lists by allowing selected third parties to use these lists to send commercial emails to potential customers.

[3] In order to increase engagement with the content of such emails, such emails may be sent only to targeted recipients on the email address lists.

Summary

[4] Aspects of the present disclosure are defined in the accompanying independent claims.

Brief description of the drawings

[5] Examples of the present disclosure will now be explained with reference to the accompanying drawings in which:

Fig. 1 is a block diagram of an email demand-side platform;

Fig. 2 is a block diagram of a sending server for use with the methods described herein;

Fig. 3 is a block diagram of a tracking server for use with the methods described herein;

Fig. 4 is a block diagram of a client for use with the methods described herein; Figs. 5a and 5b are two parts of a flowchart illustrating the steps of a method of operating the sending server;

Fig. 6 is a flowchart illustrating the steps of a method of operating the client; and

Fig. 7 is a flowchart illustrating the steps of a method of operating the tracking server.

[6] Throughout the description and the drawings, like

reference numerals refer to like parts.

Detailed description

[7] In overview, a method of operating an email demand-side platform is provided. A demand-side platform may connect buyers and sellers of advertising space, the pricing of which may be determined in real-time. The method brings together suppliers of email addresses of email recipients, and parties wishing to send targeted emails to those email recipients. By enriching a profile for each email recipient, emails can be sent to only the most relevant email recipients, thereby reducing unwanted email traffic.

[8] The method acquires the behavioural data by sending a probe email to each email recipient, the probe email

initiating placement of a cookie on that client by a tracking server. Later, when a user of that client is browsing a website affiliated with the tracking server, the tracking server is able to read the cookie and record that user' s browsing behaviour against that cookie. Subsequently, the tracking server may send the data that has been collected back to the sending server so as to enrich the profiles.

[9] The method may be implemented by way of the apparatus illustrated in Fig. 1. The apparatus includes a sending server 100, one or more clients 200 (eg, 200a, 200b) , and a tracking server 300. In the following disclosure, the platform' , 'demand-side platform' or 'email demand-side platform' may refer either to only the sending server 100, or to the sending server 100 and tracking server 300.

[10] The sending server (or 'email sending server', 'email sending device' , 'probing server' , 'probing device' , 'data management platform' ) 100 initiates sending of the probe emails, builds the profiles, and sends the targeted emails. The sending server 100 may operate its own electronic mail service, or may initiate sending of emails by communicating with a separate outgoing mail server, such as a Simple Mail Transfer Protocol (SMTP) server.

[11] The clients (or 'client devices', 'email receiving devices' , 'email clients' ) 200, which are shown as 200a and 200b in Fig. 1, receive the probe emails, and may also receive the targeted emails. Each of the clients 200 may be any device that is capable of receiving emails, eg, a desktop computer, or a mobile electronic device such as a smartphone, tablet, or laptop.

[12] The tracking server (or 'data collecting server', 'data collecting device' , 'third party data server' , 'third party data company server' ) 300 initiates placement of a cookie on each of the clients, and records browsing behaviour of the users of each of those clients.

[13] The sending server 100 communicates with the one or more clients 200 via a first communications network 400a, and the tracking server 300 communicates with the one or more clients 200 via a second communications network 400b. The first and second communications networks may be a single network, eg, the Internet.

[14] A block diagram of an exemplary sending server 100 is shown in Fig. 2. The sending server 100 comprises a processor 110 arranged to execute computer-readable instructions as may be provided to the sending server 100 via one or more of a memory 120 or a network interface 130.

[15] The memory 120, for example a random access memory (RAM), is arranged to be able to retrieve, store, and provide to the processor 110, instructions and data that have been stored in the memory 120. The network interface 130 is arranged to enable the processor 110 to communicate with a communications network, such as the Internet.

[16] Similarly, a block diagram of an exemplary tracking server 300 is shown in Fig. 3. The tracking server 300 comprises a processor 310 arranged to execute computer- readable instructions as may be provided to the tracking server 300 via one or more of a memory 320, or a network interface 330.

[17] The memory 320, for example a random access memory (RAM), is arranged to be able to retrieve, store, and provide to the processor 310, instructions and data that have been stored in the memory 320. The network interface 330 is arranged to enable the processor 310 to communicate with a communications network, such as the Internet.

[18] A block diagram of an exemplary client 200 is shown in Fig. 4. The client 200 comprises a processor 210 arranged to execute computer-readable instructions as may be provided to the client 200 via one or more of a memory 220, a network interface 230, or an input interface 250.

[19] The memory 220, for example a random access memory (RAM), is arranged to be able to retrieve, store, and provide to the processor 210, instructions and data that have been stored in the memory 220. The network interface 230 is arranged to enable the processor 210 to communicate with a communications network, such as the Internet. The input interface 250 is arranged to receive user inputs provided via an input device (not shown) such as a mouse, a keyboard, or a touchscreen. The processor 210 is further coupled to a display adapter 240, which is in turn coupled to a display device (not shown) .

[20] Figs. 5a and 5b show a flowchart of the steps of a method of operating an email demand-side platform. The steps of this method are performed at the sending server 100.

[21] At step S105, one or more email addresses are received, each email address corresponding to a respective email

recipient. The one or more email addresses may be received from a local database, or from another server. The email addresses may be uploaded to the platform/sending server by an email address supplier.

[22] At step S110, an identifier is assigned to each of the one or more email recipients. Each identifier uniquely identifies an email recipient.

[23] Each of the identifiers may be based on an email address for the corresponding email recipient. In that case, each identifier may be assigned by hashing the corresponding email address. The hashing may be performed using a cryptographic hash function, such as Advanced Encryption Standard 256 (AES 256), Secure Hash Algorithm 1 (SHA-1) or Message-Digest

Algorithm 5 (MD5) , in which case it would be difficult to recover an email address from a corresponding (hashed) identifier. For example, if the MD5 hash function is used, an identifier of b28d5fe8da784e36235a487c03a47353 may be assigned to an email recipient having the email address

abc@ e x amp1 e . com . In this way, in the event that the

identifiers are received and processed by multiple parties, the party that is assigning the identifiers (eg, the operator of the demand-side platform) may keep the corresponding email addresses private from the other parties receiving and

processing the identifiers (eg, the operator of the tracking server 300) . Such an arrangement results in enhanced security and privacy.

[24] At step S115, the one or more identifiers are received. The one or more identifiers may be generated locally on the sending server 100 according to steps S105 to S110, or may instead be received from an email address supplier performing steps S105 to S110 on its own server.

[25] At step S120, a probe email is caused to be sent to each of the one or more email recipients. Each probe email is operable to initiate placement of a packet of data associated with the corresponding identifier on a respective client 200 by the tracking server 300. This placement may occur when the email is received, opened, read, forwarded, engaged with, and/or responded to, depending on the configuration of the respective clients 200.

[26] In particular, the probe email contains a uniform

resource locator (URL) referring to the tracking server 300, and a portion of the URL contains the corresponding

identifier. In this way, when a request is made to the URL, the tracking server 300 can associate the request with a given email recipient. As a result of the sending of the probe email, a request to the URL may be initiated, eg, when the email is opened.

[27] The URL may, for example, be a reference to an image. The image may be arranged to not be visible when the probe email is opened. As one example, the image may be a lxl pixel image, also referred to as a bracking pixel' , and may have a white background. When the URL is a reference to an image, the URL may be contained in a HyperText Markup Language (HTML) image (or ^> <img>' ) tag. Examples of such an image tag are given below, where TRACKINGSVR denotes the hostname of the tracking server 300, and IDENTIFIER denotes the identifier of the email recipient.

<img src="h ttp : //TRACKINGSVR/track . jpg?id=IDENTIFIER" />

<img src="h ttp : //TRACKINGSVR/IDENTIFIER, jpg" />

<img src="h ttp : //TRACKINGSVR/track . php ?id=IDENTIFIER" />

[28] As an alternative, the URL may be a reference to a script, and the URL may be contained in an HTML script (or ^> <script>' ) tag. Example of such a script tag are given below, where again TRACKINGSVR denotes the hostname of the tracking server 300, and IDENTIFIER denotes the identifier of the email recipient.

<script src="h ttp : //TRACKINGSVR/track . j s ?id=IDENTIFIER" />

<script src="h t tp : //TRACKINGSVR/IDENTIFIER.js" />

<script src="http: //TRACKINGSVR/track .php?id=IDENTIFIER" />

[29] When a request to the URL is initiated, the tracking server 300 may be able to place a packet of data, such as a cookie, on the respective client 200, as explained at step S215 below. The tracking server 300' s ability to place the packet of data may be subject to the client 200' s

configuration, including its security settings.

[30] The packet of data is associated with the corresponding identifier, eg, it may contain the corresponding identifier. In this way, data collected by the tracking server 300 can be associated with a specific email recipient.

[31] The probe email may contain a notice that informs the recipient that a packet of data will be placed on their device. In addition, the probe email may contain one or more ^x opt-out' links, as explained in more detail below. [32] At step S125, consequent to the placement, behavioural data is acquired from the tracking server 300 for each of the one or more email recipients using the one or more

identifiers. The acquisition of the behavioural data by the tracking server 300 is explained in more detail in Figs. 6 and 7 below.

[33] The behavioural data may be requested from the tracking server 300, or may be sent unsolicited. The behavioural data may be sent as a structured set of data, and the data may be indexed according to the one or more identifiers.

[34] The behavioural data may, for example, comprise a list of websites that have been visited by each of the email

recipients, as well as the dates and times of each visit.

[35] The behavioural data is a type of probabilistic data. Other types of probabilistic data that may be acquired include open/click data and social media data, as explained below.

[36] At step S130, consequent to the causing of step S120, open/click data is acquired for each of the one or more email recipients. Open/click data is data that relates to whether an email has been opened and/or whether one or more links contained in the email have been clicked. Open/click data may, as one example, suggest that an email recipient is interested in music, because they have opened an email about music, or because they have clicked on one or more links contained in an email about music.

[37] At step S135, deterministic data is acquired for each of the one or more email recipients. Deterministic data is factual data relating to the email recipients. Exemplary deterministic data relating to an email recipient includes a first name, surname, address, date of birth, Internet Protocol

(IP) address, telephone number, date and/or time that the email recipient agreed to share their data (eg, with the email address supplier), or transactional data. The deterministic data may be obtained from the email address supplier, and may be acquired along with the one or more email addresses or the one or more identifiers.

[38] At step S140, based on at least one of the behavioural data, the open/click data or the deterministic data, a profile of each of the one or more email recipients is built.

[39] By centralising this information in one place, a single customer view (SCV) is provided. The profiles allow the demand-side platform to target email recipients based on in- market behaviour (ie, what they may be in the market to purchase) and past purchase behaviour (ie, what they have purchased in the past) .

[40] At step S145, at least one targeting criterion is

received. The targeting criterion allows a party seeking to send a targeted email to define the email recipient (s) they wish to target. For purposes of illustration, exemplary targeting criterion might be:

• all female email recipients who are interested in sports;

• all email recipients who are interested in pensions; or

• all email recipients who are known to have purchased a bicycle and are interested in technology.

[41] At step S150, based on the at least one targeting

criterion and the profiles of the plurality of email

recipients, at least one target email recipient is selected from the plurality of email recipients. The selected target email recipients form a candidate set of email recipients.

[42] As a result, the platform is able to predictively

determine and segment which identifiers are most relevant to receive a targeted email, based on any number of targeting criteria that are input into the platform. Those criteria may be deterministic, behavioural, or probabilistic, as set out in steps S125-S135 above.

[43] Steps S145-S150 may be performed iteratively, and a number of email recipients in the candidate set of email recipients may be reported after each iteration. For example, a party wishing to send a targeted email may begin with a relatively broad criterion (or a small number of criteria) , and may then narrow the selection in order to reduce the number of target email recipients, which may reduce the cost of sending the targeted emails.

[44] A wide variety of types of profiles and targeting

criterion can be accommodated, depending on the data that is used to populate the profiles, and depending on the

requirements of parties seeking to send targeted emails.

[45] Email recipients may be excluded from the selection for any number of reasons. In particular, a suppression list comprising at least one identifier for an email recipient to which emails are not to be sent may be received, and the at least one email recipients to which emails are not to be sent may be excluded from the candidate set. Email recipients may, as one example, be excluded from the selection if those email recipients are existing customers of the party seeking to send the targeted emails. This process is known as deduplication .

[46] Email recipients may also be excluded from the selection if they Unsubscribe' or ^> opt-out' of receiving targeted emails. Such an opt-out may be a global' (or platform level' ) opt-out, or a partial' (or publisher level' ) opt- out. A global' opt-out from a particular email recipient indicates that no emails should be sent to that email

recipient by the platform. A partial' opt-out indicates that the email recipient should not receive emails from a given publisher. However, emails can still be sent to that email recipient if there is at least one more publisher with the same record.

[47] A global' opt-out may be implemented as follows. A global unsubscribe request for an email recipient to which emails are not to be sent is received. In response to

receiving the global unsubscribe request, sending of targeted emails to the email recipient to which emails are not to be sent is suppressed. The email recipient to which emails are not to be sent may be identified using an identifier

corresponding to that email recipient, and that identifier may be contained in the global unsubscribe request.

[48] A partial' opt-out may be implemented as follows. In addition to receiving the one or more identifiers in step S115, said identifiers forming part of a first set of

identifiers, a second set of one or more identifiers is received, each of those identifiers also corresponding to a respective email recipient. The second set may, for example, originate from a second, different email address supplier. In the event that the first set of identifiers is assigned by hashing corresponding email addresses, the second set of identifiers may be assigned in the same way, eg, using the same cryptographic hash function. A partial unsubscribe request for an email recipient to which emails are not to be sent is received. The unsubscribe request is associated with one of the first set of identifiers or the second set of identifiers. For example, the unsubscribe request may

identify the publisher of the email that resulted in the unsubscribe request being made.

[49] If an identifier corresponding to the email recipient to which emails are not to be sent is present in both the first and second sets of identifiers, at least one targeted email may be caused to be sent to the email recipient to which emails are not to be sent. This is because that email

recipient is present on multiple lists of email addresses, but has only opted-out of receiving emails associated with a single publisher. If, however, the identifier corresponding to the email recipient to which emails are not to be sent is present in only one of the first and second sets of

identifiers, sending of targeted emails to the email recipient to which emails are not to be sent is suppressed.

[50] At step S155, a targeted email is caused to be sent to each of the at least one target email recipients. As set out above, the selection of the at least one target email

recipients may involve excluding some email recipients. In that case, a targeted email may be sent to each of the at least one target email recipients remaining in the candidate set .

[51] The number of targeted emails that is sent may be

limited, so as to avoid over-targeting some email recipients. For example, the number of targeted emails sent to each of the at least one target email recipients within a predetermined period may be limited, eg, to a predetermined number.

[52] At step S160, the sending of the at least one targeted emails is recorded. In this way, a record of all of the sent emails is kept, eg, for auditing or accounting purposes.

[53] At step S165, the results of the sending are reported to the email address supplier. In this way, the email address supplier is kept informed as to the emails that have been sent to email addresses that they provided, eg, the number or type of the emails, and the industries/sectors of the recipients that have been targeted (the Verticals' ) . In addition, the email address supplier may be financially compensated

according to the emails that have been sent to email addresses that they provided. [54] The email demand-side platform may also be used by email address suppliers to identify target recipients from their own email address lists, assuming profiles for those target recipients already exist on the demand-side platform/sending server .

[55] In other words, once (enriched) profiles have been created and stored on the demand-side platform, companies may plug in their first party customer files, in the form of customer relationship management (CRM) data, and any matches are appended with the extra information that the platform holds on those companies' customers. Those companies can thereby better understand their customers, their interests and their in-market behaviours, which they would otherwise be blind to. As a result, those companies can send more

personalised/targeted emails.

[56] As one example, if a shoe manufacturer were about to launch a financial product, they could upload their mailing list to the platform, and the platform could then send

targeted emails to only the recipients on the mailing list who have an interest in finance.

[57] The process for enriching first party customer files is as follows. A second set of one or more identifiers is received, each of those identifiers also corresponding to a respective email recipient. In the event that the first set of identifiers is assigned by hashing corresponding email addresses, the second set of identifiers may be assigned in the same way, eg, using the same cryptographic hash function. A plurality of common identifiers that are present in both the first and second sets of identifiers is identified. At least one targeting criterion is received. Based on the at least one targeting criterion and the profiles of the email

recipients corresponding to the common identifiers, at least one target email recipient from the plurality of email recipients corresponding to the common identifiers is selected .

[58] Fig. 6 shows a flowchart of the steps of another method of operating an email demand-side platform. The steps of this method may be performed at one or more of the clients 200.

[59] At step S205, a probe email addressed to an email

recipient is received. The probe email is received from the sending server 100, optionally via an outgoing mail delivery server .

[60] At step S210, consequent to receiving the probe email, a request to the tracking server 300 is initiated. The request comprises an identifier corresponding to the email recipient. This step is explained in more detail at step S120 above.

[61] At step S215, an instruction by the tracking server 300 to place a packet of data on the client 200 is executed. The instruction may be in the form of a Hypertext Transfer

Protocol (HTTP) header instruction, or a Javascript

instruction .

[62] At step S220, a web page is accessed, the web page being associated or affiliated with the tracking server 300. For example, the web page may include a call to a script that is hosted on the tracking server 300.

[63] At step S225, consequent to accessing the web page, identifying information for the web page and at least part of the packet of data is sent to the tracking server 300. In particular, the identifier for the email recipient may be sent to the tracking server 300. In one example, the packet of data, a cookie, is read using a Javascript instruction or an HTTP header instruction. [64] Fig. 7 shows a flowchart of the steps of another method of operating an email demand-side platform. The steps of this method may be performed at the tracking server 300.

[65] At step S305, one or more requests are received. Each request comprises an identifier corresponding to a respective email recipient.

[66] As set out above in relation to step S120, each of the one or more requests contains a respective URL, a portion of which contains the corresponding identifier. The request may, for example, be a request for an image. The URL may identify a script (eg, a PHP script) that generates the image and places a packet of data on the respective client 200.

[67] At step S310, one or more packets of data associated with each of the one or more identifiers are placed on the one or more respective clients 200. Each of the one or more packets of data may contain the associated identifier, which is obtained from the URL.

[68] Once the packet of data has been placed, the tracking server 300 is able to collect browsing behaviour data while the email recipient browses the World Wide Web.

[69] At step S315, behavioural data is collected for each of the one or more email recipients using the one or more

identifiers, as set out at steps S220 to S225 above.

[70] In particular, when an email recipient is visiting a website that is associated or affiliated with the tracking server 300, eg, a website that has an agreement with the operator of the tracking server 300 and has placed a call to a script on the tracking server 300 on their website, the tracking server 300 reads the packet of data, which includes the identifier for the email recipient, and stores the

browsing history data against that identifier. [71] At step S320, the behavioural data for each of the one or more email recipients is provided as a structured set of data. The set of data may be indexed according to the one or more identifiers .

[72] The behavioural data for a given email recipient may be provided in response to a request from the sending server 100, the request including the identifier for the given email recipient .

[73] The email demand-side platform may be a ^elf-service' platform, in that parties that wish to send targeted emails can set up an email campaign without human assistance from the operator of the platform. As a result, those parties have full control over the content of the targeted emails that are sent .

[74] The packet of data may be a cookie, ie, a packet of data that is placed by a website and stored on a user' s computer by the user' s web browser, and which can then later be read by that website, and sometimes also other associated websites.

[75] The behavioural data may comprise browsing history data, ie, data that relates to webpages or websites that have been visited by a particular email recipient. The behavioural data may also comprise social media data, ie, data relating to a particular email recipient that has been acquired via social networks such as Facebook, Instagram, Linkedln, etc. The behavioural data and/or deterministic data may also comprise customer relationship management, CRM, data.

[76] The building of profiles and selection of target email recipients may be performed using a machine learning

algorithm, as will be known to a person skilled in the art.

[77] Unless otherwise indicated, the steps of the methods set out above need not all be performed, and some steps may be omitted without departing from the scope of the disclosure. Furthermore, unless otherwise indicated, the steps need not be performed sequentially, and may be performed in any order.

[78] It should be noted that the steps set out above may be performed repeatedly. In particular, data may be received by the platform/sending server 100 in real-time, thereby

continuously improving the platform' s SCV of each email recipient .

[79] The steps of the method set out above in relation to Fig. 5 need not all be performed by the same (sending) server, but could instead be distributed across multiple servers/devices. Similarly, the steps of the method set out above in relation to Fig. 7 need not all be performed by the same (tracking) server, but could also be distributed across multiple

servers/devices. Alternatively, the steps of the methods of Fig. 5 and Fig. 7 could be performed by a single server, ie, the functions of the sending server 100 and tracking server 300 could be combined.

[80] In the present disclosure, the term 'email address' may refer to any identifier that is used in a system for

exchanging digital communications between recipients and identifies one of those recipients. Similarly, the term

'email recipient' may denote any recipient of digital

communications in such a system. For example, the present disclosure may apply equally to the sending of communications to Facebook users, the users being identified by their

Facebook username.

[81] A technical effect of the present disclosure is the reduction of unnecessary network traffic. Indeed, by sending targeted emails, the sending of unwanted emails is reduced.

[82] A further technical effect of the present disclosure is the reduction of storage space needed for storing emails. Indeed, by sending targeted emails, a reduced number of emails may be sent.

[83] A further technical effect of the present disclosure is the provision of an improved man-machine interface. Indeed, by automatically collecting behavioural data, an email

recipient need not manually provide information relating to their interests in order to receive targeted emails.

[84] A further technical effect of the present disclosure is efficiently reducing the size of a set of email recipients.

[85] A method for operating an email demand-side platform is provided. The method comprises receiving one or more

identifiers, each identifier corresponding to a respective email recipient; and acquiring behavioural data for each of the one or more email recipients using the one or more

identifiers. The method may further comprise receiving at least one targeting criterion; and based on the at least one targeting criterion and the behavioural data for each of the one or more email recipients, selecting at least one target email recipient from the one or more email recipients.

[86] Those skilled in the art will recognise that a wide variety of modifications, alterations, and combinations can be made with respect to the above described examples without departing from the scope of the disclosed concepts, and that such modifications, alterations, and combinations are to be viewed as being within the ambit of the disclosed concepts.

[87] Those skilled in the art will also recognise that the scope of the invention is not limited by the examples

described herein, but is instead defined by the appended claims .

[88] The approaches described herein may be embodied on a computer-readable medium, which may be a non-transitory computer-readable medium. The computer-readable medium carries computer-readable instructions arranged for execution upon a processor so as to make the processor carry out any or all of the methods described herein.

[89] The term "computer-readable medium" as used herein refers to any medium that stores data and/or instructions for causing a processor to operate in a specific manner. Such storage medium may comprise non-volatile media and/or volatile media. Non-volatile media may include, for example, optical or magnetic disks. Volatile media may include dynamic memory. Exemplary forms of storage medium include, a floppy disk, a flexible disk, a hard disk, a solid state drive, a magnetic tape, or any other magnetic data storage medium, a CD-ROM, any other optical data storage medium, any physical medium with one or more patterns of holes, a RAM, a PROM, an EPROM, a FLASH-EPROM, NVRAM, and any other memory chip or cartridge.