WO2019011727A1

WO2019011727A1 - Upgrade recommendation engine

Info

Publication number: WO2019011727A1
Application number: PCT/EP2018/067986
Authority: WO
Inventors: Jessika MALMCRONA; Geir SJURSETH; Thomas Nilsson; Nils UNDÉN; Roger WALLENTIN
Original assignee: Ebuilder AB
Current assignee: Ebuilder AB
Priority date: 2017-07-14
Filing date: 2018-07-03
Publication date: 2019-01-17
Anticipated expiration: 2020-01-14

Abstract

In accordance with one or more embodiments herein, a system (100) for providing upgrade recommendations for users of consumer electronic devices is provided. The system (100) comprises at least one central processing arrangement (150) and a plurality of consumer electronic devices (110), each comprising data collecting software for collecting device performance data. The at least one central processing (arrangement 150) is arranged to receive device performance data collected by the data collecting software (120), analyze the device performance data, generate an upgrade recommendation engine based at least on the device performance data using machine learning, and use the generated upgrade recommendation engine to output an upgrade recommendation for a user of a consumer electronic device (110) together with an upgrade propensity P_u for the user to upgrade the consumer electronic device (110).

Description

UPGRADE RECOMMENDATION ENGINE

TECHNICAL FIELD

The present disclosure relates generally to systems and methods for providing upgrade recommendations for users of consumer electronic devices.

BACKGROUND

It is difficult for users of consumer electronic devices to make a reasoned decision as to what might be an ideal device that fits the user's needs.

US2014/0195297 describes a system that analyzes usage patterns for different functionalities of consumer electronic devices, and based on the usage statistics provides upgrade recommendations. Data is collected from the user device when a functionality of the user device is activated, and the collected data includes e.g. the points of time and duration of the use of the functionality. Using this information, the most used functions and the functions not used at all can be identified, and the ideal device for the user can be recommended based on the usage habits. US2012/0117097 describes a method of utilizing user feedback about a personal computer for providing recommendations on how to upgrade the personal computer with new software and/or hardware based on use pattern data.

PROBLEMS WITH THE PRIOR ART

The system described in US2014/0195297 can only provide upgrade recommendations in simple and straightforward situations, where it is more or less obvious that a consumer electronic device with different specifications would be more suitable for a user than the current device.

The main purpose of the method described in US2012/0117097 is improving the operation of the personal computer by upgrading it with new software and/or hardware.

There is thus a need for an improved system for providing upgrade recommendations for users of consumer electronic devices.

l SUMMARY

The present disclosure relates to systems and methods for providing upgrade recommendations for users of consumer electronic devices. Prior art systems do not utilize the possibilities enabled by the collection of device performance data from large amounts of users for providing upgrade recommendations, and they have not addressed the problem that the users will not always appreciate receiving such upgrade recommendations.

The above described problems are addressed by the claimed system for providing upgrade

recommendations for users of consumer electronic devices. The system may comprise at least one central processing arrangement and a plurality of consumer electronic devices, each comprising data collecting software for collecting device performance data. The at least one central processing arrangement may be arranged to: receive device performance data collected by the data collecting software from the plurality of consumer electronic devices; analyze the device performance data; generate an upgrade recommendation engine based at least on the device performance data using machine learning; and use the generated upgrade recommendation engine to output an upgrade recommendation for a user of a consumer electronic device, together with an upgrade propensity for the user to upgrade the consumer electronic device. This enables the system to provide a very accurate upgrade recommendation for a user of a consumer electronic device, at a point in time when such an upgrade recommendation is likely to be appreciated by the user.

In embodiments, the system comprises a plurality of consumer electronic devices of the same or similar model, wherein information regarding upgrade actions taken by the users of the consumer electronic devices is also transferred to the at least one central processing arrangement, and the upgrade recommendation engine is generated based also on upgrade actions taken by users of consumer electronic devices having the same or similar model experiencing similar device performance data. This improves the upgrade recommendation engine.

The above described problems are further addressed by the claimed method for providing upgrade recommendations for users of consumer electronic devices. The method may comprise: collecting device performance data from a plurality of consumer electronic device using data collecting software; transferring device performance data collected by the data collecting software to at least one central processing arrangement; analyzing the device performance data; generating an upgrade recommendation engine based at least on the device performance data using machine learning; and providing, as an output from the generated upgrade recommendation engine, an upgrade recommendation for a user of a consumer electronic device together with an upgrade propensity for the user to upgrade the consumer electronic device. This enables the providing of very accurate upgrade recommendations, at times when such upgrade recommendations are likely to be appreciated by the users. In embodiments, device performance data is collected from a plurality of consumer electronic devices of the same or similar model, and the method further comprises transferring information regarding upgrade actions taken by the users of the consumer electronic devices to the at least one central processing arrangement. The generating of the upgrade recommendation engine may then be based also on upgrade actions taken by users of consumer electronic devices having the same or similar model experiencing similar device performance data. This improves the upgrade recommendation engine.

In embodiments, the upgrade recommendation comprises suitable technical specifications and/or suitable model of a new consumer electronic device for a user of a consumer electronic device. This enables a user to find an ideal device that fits the user's needs. In embodiments, the upgrade actions include upgrading the consumer electronic device to a new model, and the upgrade recommendation comprises a suitable model based on the model selected by users of consumer electronic devices having the same or similar model and experiencing similar device performance data. This improves the upgrade recommendation.

In embodiments, the data collecting software is installed in each of the plurality of consumer electronic devices by being downloaded as part of an application.

In embodiments, the central processing arrangement is arranged to also determine whether to use the generated upgrade recommendation engine to output an upgrade recommendation for a user of a consumer electronic device based on a comparison of the upgrade propensity for the user to upgrade the consumer electronic device with at least one at least one upgrade propensity threshold for the user of the consumer electronic device.

In embodiments, the device performance data includes the charging pattern, the average battery life, and/or the state of health (SOH) of a power source such as a battery comprised in the at least one consumer electronic device.

In embodiments, the device performance data includes the free storage space in a storage memory of the consumer electronic device.

The consumer electronic device may e.g. be a portable communications device, such as e.g. a smartphone.

The plurality of consumer electronic devices may e.g. be hundreds, or thousands, or millions of consumer electronic devices.

The at least one central processing arrangement may be one central processing arrangement, or a number of central processing arrangements between which signals are transmitted. Some processing may e.g. take place in one central processing arrangement, and signals may then be transmitted to one or more other central processing arrangements for further processing.

The scope of the invention is defined by the claims, which are incorporated into this section by reference. A more complete understanding of embodiments of the invention will be afforded to those skilled in the art, as well as a realization of additional advantages thereof, by a consideration of the following detailed description of one or more embodiments. Reference will be made to the appended sheets of drawings that will first be described briefly.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 schematically illustrates a system for providing upgrade recommendations for users of consumer electronic devices, in accordance with one or more embodiments described herein.

Figure 2 is a schematic conceptual overview of a system for providing upgrade recommendations for users of consumer electronic devices, in accordance with one or more embodiments described herein.

Figure 3 is a schematic component overview of a system for providing upgrade recommendations for users of consumer electronic devices, in accordance with one or more embodiments described herein. Figure 4 is a schematic overview of an example process and upgrade propensity for an example user of a consumer electronic device, in accordance with one or more embodiments described herein.

Figure 5 schematically illustrates a method for providing upgrade recommendations for users of consumer electronic devices, in accordance with one or more embodiments described herein.

Embodiments of the present disclosure and their advantages are best understood by referring to the detailed description that follows. It should be appreciated that like reference numerals are used to identify like elements illustrated in one or more of the figures.

DETAILED DESCRIPTION

Since the number of consumer electronic devices such as e.g. mobile telephones in the world is very large, many conclusions may potentially be drawn from an analysis of data, such as e.g. device performance data, from these consumer electronic devices. If such data could be properly analyzed, there is a potential for creating a very accurate upgrade recommendation engine that could recommend a suitable consumer electronic device to a user based on these data. The data may e.g. be collected by data collecting software in the form of a device inspector, which is a specially developed software for data collection that may be added to any application, regardless of who has created the application. Such applications may be provided by anyone wishing to collect data from consumer devices, and they may be provided directly to users for downloading, without involving any device manufacturers.

The present disclosure relates generally to systems and methods for providing upgrade recommendations for users of consumer electronic devices. Embodiments of the disclosed solution are presented in more detail in connection with the figures.

Figure 1 schematically illustrates a system 100 for providing upgrade recommendations for users of consumer electronic devices 110, in accordance with one or more embodiments described herein. The system 100 may comprise at least one central processing arrangement 150 and a plurality of consumer electronic devices 110, each comprising data collecting software 120 for collecting device performance data. The plurality of consumer electronic devices 110 may e.g. be hundreds, or thousands, or millions of consumer electronic devices 110. Device performance data collected by the data collecting software 120 may be received in the at least one central processing arrangement 150, where the device performance data may be analyzed and an upgrade recommendation engine may be generated based at least on the device performance data using machine learning. The data collecting software 120 may e.g. be a device inspector integrated into at least one application that has been downloaded into each of the consumer electronic devices 110. Each of the computer electronic devices may further comprise at least one battery 130 and at least one storage memory 140.

Figure 2 is a schematic conceptual overview of a system 100 for providing upgrade recommendations for users of consumer electronic devices 110, in accordance with one or more embodiments described herein. A central processing arrangement 150 receives device data, e.g. device performance data, from a community 200 of consumer electronic devices 110. The community 200 may e.g. comprise hundreds, or thousands, or millions of consumer electronic devices 110. The central processing arrangement 150 analyses the data and generates an upgrade recommendation engine based at least on the device performance data using machine learning. Machine learning enables the system 100 to handle device performance data from huge amounts of consumer electronic devices 110. The upgrade recommendation engine may output upgrade recommendations to users of consumer electronic devices 110 in the community 200. The upgrade recommendation engine may also output upgrade propensities P_u for the users in the community 200 to upgrade their consumer electronic devices 110. The central processing arrangement 150 may also receive other forms of data, such as user data and/or upgrade actions, from the community 200 of consumer electronic devices 110. The upgrade

recommendation engine may be generated based also on these data, preferably using machine learning.

Figure 3 is a schematic component overview of a system for providing upgrade recommendations for users of consumer electronic devices 110, in accordance with one or more embodiments described herein. A device inspector 120 locally running in a consumer electronic device 110 periodically collects device information, such as e.g. device performance data, and transmits this information, e.g. as an "event", to a central processing arrangement 150, where the data can be analyzed. The data may e.g. be device performance data and/or information about upgrade actions. Data may be collected from a whole community 200 of consumer electronic devices 110. The collected data may e.g. be used for generating an upgrade recommendation engine. The upgrade recommendation engine may be arranged to output upgrade recommendations for users of consumer electronic devices 110 in the community 200. The upgrade recommendations may e.g. comprise suitable technical specifications and/or suitable model of a new consumer electronic device 110. Users of the community 200 of consumer electronic devices 110 may take upgrade actions, such as upgrading their consumer electronic device 110 to a new model. Information about such upgrade actions may be transferred to the central processing arrangement 150, where it may contribute to the generation or training of the upgrade recommendation engine.

The central processing arrangement 150 may also be arranged to output an upgrade propensity P_u for a user of a consumer electronic device 110. The upgrade propensity P_u may be defined as the likelihood that the user upgrades the consumer electronic device 110. The upgrade propensity P_u is therefore a value that changes continuously - the likelihood that a user upgrades the consumer electronic device 110 at any point of time in the future is of course almost 100%, but the likelihood that a user upgrades the consumer electronic device 110 at a given point of time is more difficult to determine. The upgrade propensity P_u may be a general propensity for upgrading the consumer electronic device 110, or a propensity for upgrading to a specific model of the consumer electronic device 110.

The upgrade propensity P_u may e.g. be determined using a predictive model. The iterative process to establish a predictive model for an upgrade propensity P_u may e.g. be implemented using supervised regression analysis, e.g. by optimizing a multinomial regression algorithm such as Pu(t) = w_a ^*a + Wb^*b + Wc^*c + Wd^*d + w_e ^*e + Wf^*f, where the weights w_a-Wf may be calculated based on training data received form the community 200 of consumer electronic devices 110, e.g. using machine learning. The predictive model for calculating the upgrade propensity P_u may also be determined entirely by machine learning based on data received from the community 200 of consumer electronic devices 110, without using any predetermined algorithm format.

Training of the predictive model may include using a training data set which binds specific data points to actual upgrade events in order to perform a regressions analysis. Upgrade events may be automatically tracked by monitoring users and devices in the community 200 of consumer electronic devices 110 and specifically capturing all changes of IMEIs (device identifiers) relative to an individual subscriber identity. Subscribers may e.g. identified by capturing the IMSI (international mobile subscriber identity) from the device. As upgrade events are captured in the community 200, the associated device data from those consumer electronic devices 110 may be used to construct labeled features that train the predictive model and keep it continuously up to date to changing trends in user behavior.

The predictive model for calculating the upgrade propensity P_u may also represent different measures of experience impediments to device users, since as experience deteriorates the upgrade propensity P_u increases. Experience impediments may be derived from continuously sampled device performance data from individual devices on a regular time interval. The upgrade propensity P_u may further depend on user data such as e.g. assessed income level, age, gender, personal interests, subscription contract, etc., which may be assessed e.g. from monitoring and analyzing device data in terms of device features or applications actually used by the users in the community 200 of consumer electronic devices 110.

Predictive models of the above described kinds may also be used for determining other types of upgrade recommendations, such as e.g. suitable technical specifications and/or suitable model of a new consumer electronic device 110. The central processing arrangement 150 may e.g. continuously scan the data and use machine learning updates, community propensities, and individual propensities for the device type and the specific user respectively, and generate an upgrade recommendation engine based on these data, e.g. using a predictive model. At any given point in time, the upgrade recommendation engine may provide an upgrade recommendation for a user of a consumer electronic device 110 based on the combination of community and individual data and propensities. The process may feed the system with individual device data for the consumer electronic device 110 and generate an upgrade recommendation engine, e.g. in the form of a descriptive analytical model. This upgrade recommendation engine may feed predictive analytics that may ultimately make an upgrade recommendation.

The upgrade recommendation engine may thus be continuously generated, or trained, based on various types of received data, using machine learning, so that it may provide useful upgrade recommendations to users of the community 200 of consumer electronic devices 110. Figure 4 is a schematic overview of an example process and upgrade propensity for an example user of a consumer electronic device 110, in accordance with one or more embodiments described herein. The upgrade propensity P_u is calculated for a user of a consumer electronic device 110 based on data received from the community 200 of consumer electronic devices 110. The upgrade propensity P_u is in figure 4 shown in a diagram together with an upgrade propensity threshold Pt determined by the central processing arrangement 150 based on data received from the community 200 of consumer electronic devices 110. When the upgrade propensity P_u is higher than the upgrade propensity threshold Pt, the user is likely to wish to upgrade the consumer electronic device 110. The user may then appreciate if the upgrade

recommendation engine provides an upgrade recommendation, either for suitable technical specifications or for a suitable model of consumer electronic device 1 10.

It may appear to be more likely that a user appreciates an upgrade recommendation the higher the upgrade propensity P_u of the user is, but this is not necessarily the case. If the upgrade propensity P_u is very high, the user has probably already made a decision to upgrade, and may be well on the way towards going through with the upgrade. An upgrade recommendation at this point may just cause irritation from the user. It may therefore be useful to use also an upper threshold for the upgrade propensity P_u, in the form of an upper upgrade propensity threshold Pt_u. This means that it is determined whether the upgrade propensity P_u is within a predetermined range between a lower upgrade propensity threshold Pti and an upper upgrade propensity threshold Ptu, rather than just above an upgrade propensity threshold Pt.

The at least one upgrade propensity threshold, that may be used in order to determine whether to use the generated upgrade recommendation engine to provide an upgrade recommendation, may be determined in various ways. In one example, the upgrade recommendation engine initially outputs an upgrade recommendation to all users of consumer electronic devices, regardless of their upgrade propensity P_u. Data can then be collected regarding whether the upgrade recommendation is appreciated by the users, e.g. by determining which users actually perform an upgrade based on the upgrade recommendation. The result of such a data collection is likely to follow a bell shaped curve, with users having low or very high upgrade propensities P_u less likely to upgrade than users having medium high upgrade propensities P_u. Based on the data, at least one upgrade propensity threshold Pt may be determined.

If the at least one upgrade propensity threshold Pt is determined in the form of a lower upgrade propensity threshold Pti and an upper upgrade propensity threshold Ptu, these thresholds may then be used to determine whether the upgrade propensity P_u is within the predetermined range between the lower upgrade propensity threshold Pti and the upper upgrade propensity threshold Ptu, and the upgrade recommendation engine may in this case be used to provide upgrade recommendations only to users having upgrade propensities P_u within this predetermined range. Example process

Different types of device performance data in the form of e.g. CPU load average, internal storage utilization of a storage memory 140, charging pattern of battery 130, average battery life of battery 130, and/or state of health (SOH) of battery 130, may be collected from a consumer electronic device 110 (a method of determining a state of health of a power source of a portable device is e.g. described in US2015/0241515). Other types of device data such as phone type category (e.g. budget/premium), phone age, amount of stored pictures, types of installed applications, NFC status uptime, and/or data and WiFi usage may also be collected. The device performance data is preferably analyzed in at least one central processing arrangement 150, and an upgrade recommendation engine is preferably generated based on different types of device performance data, possibly also in combination with other types of data. However, in this example, the device performance data for simplicity relates solely to internal storage utilization of a storage memory 140 of the consumer electronic device 110.

For the purposes of this example, we will assume that six subsets of data (D1 , D2, D3, D4, D5, D6) are collected from the consumer electronic device 110, e.g. in the form of "events", and sent to the central processing arrangement 150 for processing and analysis.

Collection of the data

Data may e.g. be collected from a consumer electronic device 110 using a locally running device inspector 120. Example of collected data in the form of an "event":

"Brand": Samsung

"HapticFeedback": "true"

"OSBuildNumber": "NRD90M. G955FXXU1AQEB "

"OSSecurityPatch": "2017-05-01"

"OSVersionSDK": "24"

"Manufacturer": "Samsung"

"OS": "Android"

"ModellD": "SM-G935F"

"OSVersionRelease": "7.0"

"ScreenTimeout": "600"

"Brightness": "84"

"IsLiveWallpaper": "false"

"ExternalUserld": "AEF-13234-234234-S32JL23JH3"

"OSVersionlncremental' "G955FXXU1AQEB " "BatteryPresent": "true"

"BatteryVoltage": "3836"

"BatteryTechnology": "Li-ion"

"BatteryPlugged": "0"

"Battery Scale": "100"

"BatteryNowCurrent": "-29"

"BatteryLevel": "50"

"BatteryHealth":

"BatteryAvgCurrent": "-43"

"BatteryStatus": "3"

"BatteryTemp": "269"

"DownMobilTraffic": "0"

"UpWifiTraffic": "31469637"

"NFC": "true"

"Uptime": "70407842"

"WifiHotspot": "false"

"InternetNetwork": "WIFI"

"UpMobilTraffic": "0"

"DownWifiTraffic": "205288316"

"CarrierO": "Turk Telekom \ 3"

"Bluetooth": "false"

"InstalledApps": "com. sec. android, app. voicenote "

"com.samsung.oh"

"com. andrewgarnson. dummydefense "

"se.hemnet.android"

"1MB": "35435808368xxxx"

"Channel": "DDCSDK"

"EventType": "Service"

"freelntStorage": "978693888"

"totallntStorage": "1530595328"

"Dimension": "150.9 x 72.6 x 7.7 mm"

"Build Fingerprint": "samsung/hero2ltexx/hero2lte:7.0/NRD90M/G935FXXU1DQE7:user/release-keys"

"Bootloader": "G935FXXU1DQE7"

"Java VM": "ART 2.1.0"

"Display Size": "5.5 inches"

"Display Resolution": "Full HD, 1080x1920 pixels" "Display Pixel Density": '534 ppi"

"Display Software Density": '480 dpi (xxhdpi)"

"DisplayRefresh Rate": '59 Hz"

"CPU Architecture": ARMv8-A"

"Board": 'universal8890"

"Chipset": 'SAMSUNGEXYNOS8890"

"Cores": '8"

"Clock Speed": '442 MHz - 2600 MHz"

"Instruction Sets": 'arm64-v8a"

"CPU Features": 'fp asimd evtstrm aes pmuii shal sha2 crc32"

"CPU Governor": Interactive"

"Kernel Version": '3.18.14-11104523"

"Kernel Architecture": 'aarch64"

"OpenGL Version": OpenGL ES 3.2"

"sim": 'Nano-SIM"

"Camera": Ί2.2 MP"

"CameraFocal Length": '4.2"

"Camera Focus Mode": Auto"

"Camera Horiz View Angle": '65.0"

"Camera Max Exp Comp": '20"

"Camera Max Exp Comp": '-20"

"Camera Zoom Supported": True"

"Flash": 'Yes"

"Secondary Camera": '5MP"

"Bluetooth": 'Yes"

"GPS": 'Yes"

"NFC": 'Yes

Analysis of the data

The internal storage utilization (freelntStorage) of the consumer electronic device 110 isolated from the rest of the data and spread over time may e.g. be as follows:

D1 : "freelntStorage": "978693888"

D2: "freelntStorage": "319014400"

D3: "freelntStorage": "891901441 "

D4: "freelntStorage": "159014320" D5: "freelntStorage": "711901472"

D6: "freelntStorage": "92190145"

As is clear from the progression of the data, there is a "jitter" effect, where the consumer electronic device 110 is periodically cleaning and recovering from the low space state and then filling up to the triggering point again. Furthermore, focusing on D2, D4, and D6, it is clear that the internal storage utilization is increasing over time for this consumer electronic device 110. This information may be used by the predictive model, so that it can predict when the user will ultimately run out of internal storage space and may desire an upgrade.

Tracking the user and devices over time

This example relies on freelntStorage to make its predictions, but ExternalUserld is also an attribute in a typical event. Assuming that the user has been tracked for a longer period of time, the system may be able to track the user's previous upgrades: e.g. from a Galaxy S3 to a Galaxy S5, and then from a Galaxy S5 to a Galaxy S7 (the user's current device). This historical data provides additional descriptive analytical data as input to the weighting model. This constant flow of new descriptive analytical data drives ongoing training for the predictive model. Community data and device tracking over time

The events may also be processed and stored for community weighting. The community data shows trends and may be used to dynamically assign weights in the predictive model. Based on trend data, the central processing arrangement 150 may e.g. have assigned a weight of 6 to any device with less than 100 MB of free space. By tracking the ExternalUserld (see the sample data above), the system can track upgrade trends for users and apply weights accordingly. Last month, the calculated weight applied to a space below a given threshold may have been 5, but today's data may make that a 7. Moreover, the thresholds themselves follow their own trends and are calculated as part of the community data weighting.

User provided feedback

Additional weighting may come in the form of user provided feedback. This data may describe for example why a user chose to upgrade, or even their own inclination toward an upgrade now and over time.

The predictive model may recognize that those users who both have a high score with regard to their upgrade propensity and have actively expressed an intention to upgrade are even more likely to perform an upgrade. This data over time is also interesting from a community trend perspective with regard to the descriptive analytical model. Upgrade propensity

Based on the total- and free-storage attributes from the data, the central processing arrangement 150 may apply weights in an ongoing fashion. Assuming that the upgrade recommendation engine is queried right after D6, the system may have the following data points and weights:

· A weight of 6 with regard to the current storage of the device

• However, the predictive model can tell us that in ten days time this will increase to an 8

• A weight of 8 regarding the jitter effect perceived over time

• Community trend data provides a weight of 7 for Samsung users with storage problems to upgrade to a newer model Samsung device

· The user previously provided feedback indicating their interest in an upgrade to solve their storage issues in the next six months and this yields a weight of 6

The upgrade recommendation engine may based on this data output an upgrade recommendation and an upgrade propensity P_u, e.g. by the predictive model predicting both when an individual is most inclined to upgrade to a new consumer electronic device 110, and which model best suits the user's needs. The descriptive analytical data may e.g. show that this user has already made several upgrades within the Samsung family in the past, and that community trend data follows this same trend. Furthermore, additional user feedback may also indicate inclination for an upgrade.

Given the community data and weight applied to the user device, the upgrade recommendation engine may e.g. output a 6 out of 10 upgrade propensity P_u that this user will upgrade to a Samsung Galaxy S8 (immediately following D6 from the data above). The predictive model may also show that the upgrade propensity P_u will likely increase to an 8 in ten days based on an analysis of internal storage utilization over time, plus community trend analysis.

This upgrade propensity P_u represented one example based on a single vector: internal storage utilization. Obviously, the upgrade propensity P_u based on combining a number of different attributes from community device trends with individual device attributes creates an aggregated score. Using machine learning, enormous amounts of data may be combined in order to generate a very accurate upgrade recommendation engine.

For all events where a user has made an upgrade or given positive feedback regarding intention of upgrading, the algorithm may be updated based on machine learning. Method embodiments

Figure 5 schematically illustrates a method 500 for providing upgrade recommendations for users of consumer electronic devices, in accordance with one or more embodiments described herein. The method 500 may comprise: Step 520: Collecting device performance data from a plurality of consumer electronic devices 110 using data collecting software 120.

Step 530: Transferring device performance data collected by the data collecting software 120 to at least one central processing arrangement 150.

Step 550: Analyzing the device performance data. Step 560: Generating an upgrade recommendation engine based at least on the device performance data using machine learning.

Step 580: Providing, as an output from the upgrade recommendation engine, an upgrade recommendation for a user of a consumer electronic device 110 and an upgrade propensity P_u for the user to upgrade the consumer electronic device 110. In embodiments, the upgrade recommendation comprises suitable technical specifications and/or suitable model of a new consumer electronic device 110 for a user of a consumer electronic device 110

The device performance data may be collected from a plurality of consumer electronic devices 110 of the same or similar model. The method 500 may then further comprise:

Step 540: Transferring information regarding upgrade actions taken by the users of the consumer electronic devices 110 to the at least one central processing arrangement 150.

The generating 560 of the upgrade recommendation may then based also on upgrade actions taken by users of consumer electronic devices 110 having the same or similar model experiencing similar device performance data.

In embodiments, the upgrade actions include upgrading the consumer electronic device 110 to a new model. The upgrade recommendation may then comprise a suitable model based on the model selected by users of consumer electronic devices 110 having the same or similar model and experiencing similar device performance data.

In embodiments, the method 500 further comprises at least one of the following: Step 510: installing the data collecting software 120 in each of the plurality of consumer electronic devices 110 by downloading it as part of an application.

Step 570: determining whether to provide 580, as an output from the upgrade recommendation engine, an upgrade recommendation for a user of a consumer electronic device 110, based comparing the upgrade propensity P_u for the user to upgrade the consumer electronic device 110 with at least one at least one upgrade propensity threshold for the user of the consumer electronic device 110.

In embodiments, the device performance data includes the charging pattern, the average battery life, and/or the state of health (SOH) of a power source such as a battery 130, comprised in the at least one consumer electronic device 110. In embodiments, the device performance data includes the free storage space in a storage memory of the at least one consumer electronic device 110.

In embodiments, the consumer electronic device 110 is a portable communications device, such as e.g. a smartphone.

The foregoing disclosure is not intended to limit the present invention to the precise forms or particular fields of use disclosed. It is contemplated that various alternate embodiments and/or modifications to the present invention, whether explicitly described or implied herein, are possible in light of the disclosure. For example, the disclosure describes smartphones as examples of consumer electronic devices 110, but the disclosure relates to other types of portable communication/telecommunication devices, or indeed any other types of consumer electronic devices. Accordingly, the scope of the invention is defined only by the claims.

Claims

1. System (100) for providing upgrade recommendations for users of consumer electronic devices (110), the system comprising at least one central processing arrangement (150) and a plurality of consumer electronic devices (110), each comprising data collecting software (120) for collecting device performance data, wherein the at least one central processing arrangement (150) is arranged to:

receive device performance data collected by the data collecting software (120) from the plurality of consumer electronic devices (110);

analyze the device performance data;

generate an upgrade recommendation engine based at least on the device performance data using machine learning; and

use the generated upgrade recommendation engine to output an upgrade recommendation for a user of a consumer electronic device (110) together with an upgrade propensity (P_u) for the user to upgrade the consumer electronic device (110).

2. System (100) according to claim 1 , wherein the upgrade recommendation comprises suitable technical specifications and/or suitable model of a new consumer electronic device (110) for a user of a consumer electronic device (110) that is a part of the system (100).

3. System (100) according to claim 1 or 2, comprising a plurality of consumer electronic devices (110) of the same or similar model, wherein information regarding upgrade actions taken by the users of the consumer electronic devices (110) is also transferred to the at least one central processing arrangement (150), and the upgrade recommendation engine is generated based also on upgrade actions taken by users of consumer electronic devices (110) having the same or similar model experiencing similar device performance data.

4. System (100) according to claim 3, wherein the upgrade actions include upgrading the consumer electronic device (110) to a new model, and the upgrade recommendation comprises a suitable model based on the model selected by users of consumer electronic devices (110) having the same or similar model and experiencing similar device performance data.

5. System according to any one of claims 1-4, wherein the data collecting software (120) is installed in each of the plurality of consumer electronic devices (110) by being downloaded as part of an application.

6. System according to any one of claims 1-5, wherein the central processing arrangement (150) is arranged to determine whether to use the generated upgrade recommendation engine to output an upgrade recommendation for a user of a consumer electronic device (110) based on a comparison of the upgrade propensity (P_u) for the user to upgrade the consumer electronic device (110) with at least one upgrade propensity threshold (Pt) for the user of the consumer electronic device (110).

7. System (100) according to any one of claims 1-6, wherein the device performance data includes the charging pattern, the average battery life, and/or the state of health (SOH) of a power source such as a battery (130) comprised in the at least one consumer electronic device (110).

8. System (100) according to any one of claims 1-7, wherein the device performance data includes the free storage space in a storage memory (140) of the at least one consumer electronic device (110).

9. System (100) according to any one of claims 1-8, wherein the consumer electronic device (110) is a portable communications device, such as e.g. a smartphone.

10. Method (500) for providing upgrade recommendations for users of consumer electronic devices, comprising:

collecting (520) device performance data from a plurality of consumer electronic devices (110) using data collecting software (120);

transferring (530) device performance data collected by the data collecting software (120) to at least one central processing arrangement (150);

analyzing (550) the device performance data;

generating (560) an upgrade recommendation engine based at least on the device performance data using machine learning; and

providing (580) as an output from the generated upgrade recommendation engine, an upgrade recommendation for a user of a consumer electronic device (110) and an upgrade propensity (P_u) for the user to upgrade the consumer electronic device (110).

11. Method (500) according to claim 10, wherein the upgrade recommendation comprises suitable technical specifications and/or suitable model of a new consumer electronic device (110) for a user of a consumer electronic device (110).

12. Method (500) according to claim 10 or 11 , wherein device performance data is collected from a plurality of consumer electronic devices (110) of the same or similar model, and the method (500) further comprises transferring (540) information regarding upgrade actions taken by the users of the consumer electronic devices (110) to the at least one central processing arrangement (150), and wherein the generating (560) of the upgrade recommendation is based also on upgrade actions taken by users of consumer electronic devices (110) having the same or similar model experiencing similar device performance data.

13. Method (500) according to claim 12, wherein the upgrade actions include upgrading the consumer electronic device (110) to a new model, and the upgrade recommendation comprises a suitable model based on the model selected by users of consumer electronic devices (110) having the same or similar model and experiencing similar device performance data.

14. Method (500) according to any one of claims 10-13, further comprising installing (510) the data collecting software (120) in each of the plurality of consumer electronic devices (110) by downloading it as part of an application.

15. Method (500) according to any one of claims 10-14, further comprising determining (570) whether to provide (580), as an output from the generated upgrade recommendation engine, an upgrade

recommendation for a user of a consumer electronic device (110), based comparing the upgrade propensity (Pu) for the user to upgrade the consumer electronic device (110) with at least one at least one upgrade propensity threshold (Pt) for the user of the consumer electronic device (110).

16. Method (500) according to any one of claims 10-15, wherein the device performance data includes the charging pattern, the average battery life, and/or the state of health (SOH) of a power source such as a battery (130) comprised in the at least one consumer electronic device (110).

17. Method (500) according to any one of claims 10-16, wherein the device performance data includes the free storage space in a storage memory (140) of the at least one consumer electronic device (110).

18. Method (500) according to any one of claims 10-17, wherein the consumer electronic device (110) is a portable communications device, such as e.g. a smartphone.