Method and apparatus for generating predictive insights for authoring short messages

1. A computer-implemented method for generating predictive insights for electronic messages, the method comprising:

receiving a message to be sent as an input;

in response to receiving the message to be sent, generating a plurality of feature vectors for the message to be sent;

accessing feature vectors for messages sent in the past and observed key performance indicator (KPI) values for the messages sent in the past;

based on an analysis of the feature vectors and the observed KPI values for the messages sent in the past, determining KPI contributions, wherein the KPI contributions comprise a measure of contribution of a feature vector to a KPI value;

predicting a KPI value for the message to be sent by applying the determined KPI contributions to the plurality of feature vectors of the message to be sent; and

presenting the predicted KPI value and the message to be sent.

2. The method as claimed in claim 1 , wherein the message to be sent is at least one of:

a subject line of a marketing message; or

a push message.

3. The method as claimed in claim 1 , wherein the plurality of feature vectors of the message to be sent area subset of the feature vectors of messages sent in the past.

4. The method as claimed in claim 1 , further comprising generating the plurality of feature vectors for the message to be sent by:

passing the message to be sent through a language parser.

5. The method as claimed in claim 1 and further comprising:

identifying one or more filters associated with messages sent in the past;

clustering messages sent in the past based on the one or more filters, wherein the messages in a given cluster have common filters; and

for each cluster, determining KPI contributions of feature vectors of messages in that cluster.

6. The method as claimed in claim 5 , wherein determining the KPI contributions comprises:

identifying a subset of the one or more filters based on the input; and

determining KPI contributions using a cluster of messages that satisfies the subset.

7. The method as claimed in claim 5 and further comprising:

predicting different values of KPI for the message to be sent by applying KPI contributions specific to each cluster to the plurality of feature vectors; and

recommending filters of a cluster for which value of KPI meets conditions specified by a marketer or for which value of the KPI is maximum.

8. The method as claimed in claim 1 and further comprising:

identifying words, in the message to be sent, that impact the predicted KPI value; and

indicating the identified words along with the impact.

9. The method as claimed in claim 8 and further comprising:

altering the message to be sent based on the impact of the identified words; and

predicting a KPI value for the altered message.

10. The method as claimed in claim 1 and further comprising:

tracking an actual KPI value for the message to be sent after the message is sent; and

using tracked KPI value as feedback to improve future predictions of KPI values.

11. A computer-implemented method for generating predictive insights for electronic messages, the method comprising:

tracking messages sent in past;

tracking values of key performance indicators (KPIs) for messages sent in past;

clustering messages based on one or more filters associated with the messages;

for each cluster, determining a predictive model for predicting KPI value using the messages in the cluster and tracked KPI values;

receiving a message to be sent as an input;

identifying a subset of the one or more filters from the input, wherein the subset is associated with the message to be sent;

selecting a predictive model from the determined predictive models using the subset of the one or more filters from the input;

predicting a KPI value for the message to be sent by applying the selected predictive model on the message to be sent;

identifying words, in the message to be sent, that impact the predicted KPI value; and

indicating the identified words along with the impact.

12. The method as claimed in claim 11 , wherein determining the predictive model for each cluster comprises:

generating feature vectors of the messages in the cluster; and

for each feature vector, determining a KPI contribution using the tracked KPI values of the messages in the cluster and the feature vectors of the messages in the cluster, wherein the KPI contribution is a measure of contribution of a feature vector to KPI value.

13. The method as claimed in claim 12 , wherein predicting the KPI value of the message to be sent comprises:

generating a plurality of feature vectors of the message to be sent; and

applying the KPI contributions determined using the tracked KPI values of the messages in the cluster and the feature vectors of the messages in the cluster to the plurality of feature vectors of the message to be sent, wherein the plurality of feature vectors of the message to be sent is a subset of feature vectors of the messages in the cluster.

14. The method as claimed in claim 12 and further comprising:

identifying words, in the message to be sent, which correspond to feature vectors that positively or negatively impact the KPI value, wherein the impact of a feature vector is determined based on corresponding KPI contribution; and

indicating the identified words along with the impact.

15. The method as claimed in claim 14 and further comprising:

altering the message to be sent based on the impact of the identified words; and

predicting KPI value for the altered message.

16. The method as claimed in claim 11 and further comprising:

tracking actual KPI value for the message to be sent after the message is sent; and

using tracked KPI value as feedback to improve the selected predictive model to improve future predictions of KPI value.

17. An apparatus for generating predictive insights for messages, the apparatus comprising:

a memory; and

one or more processors electronically coupled to the memory, the one or more processors, in conjunction with the memory, cause the apparatus to:

receive a message to be sent as an input,

access feature vectors for messages sent in past, and observed key performance indicator (KPI) values for the messages sent in past,

based on an analysis of the feature vectors and the observed KPI values for the messages sent in the past, determine KPI contributions, wherein the KPI contributions comprise a measure of contribution of a feature vector to a KPI

predict a KPI value for the message to be sent by applying the determined KPI contributions to the feature vectors of the message to be sent; and

present the predicted KPI value and the message to be sent.

18. The apparatus as claimed in claim 17 , wherein determining the KPI contributions comprises:

identifying one or more filters associated with the messages sent in the past;

clustering messages sent in the past based on the one or more filters, wherein the messages in a given cluster have common filters; and

for each cluster, determining KPI contributions of feature vectors of messages in that cluster.

19. The apparatus as claimed in claim 18 , wherein predicting the KPI value of the message to be sent comprises:

generating a plurality of feature vectors of the message to be sent; and

applying the KPI contributions, determined using the tracked KPI values of the messages sent in the past that satisfy the one or more filters and the feature vectors of the messages sent the in past that satisfy the one or more filters, to the plurality of feature vectors of the message to be sent, wherein the plurality of feature vectors of the message to be sent is a subset of feature vectors of the messages sent in the past that satisfy the one or more filters.

20. The apparatus as claimed in claim 18 wherein the one or more processors, in conjunction with the memory, cause the apparatus to:

identify words, in the message to be sent, which correspond to feature vectors that positively or negatively impact the KPI value, wherein the impact of a feature vector is determined based on a corresponding KPI contribution; and

indicate the identified words along with the impact.