Intelligent asset suggestions based on both previous phrase and whole asset performance
Example embodiments of the present disclosure provide for an example method including obtaining data indicative of user input via a user interface associated with a construction workflow for generating customized content items. The example method includes determining one or more key terms associated with the user input. The example method includes determining one or more suggested content items based on the one or more key terms. The example method includes generating a predicted performance increase associated with each respective content item of the one or more suggested content items. The example method includes providing, to a first user device, data indicative of a structured input interface associated with the construction workflow configured with one or more input fields and the one or more suggested content items, wherein the structured input interface is configured for rendering via a graphical user interface.
1. A computer implemented method comprising:
obtaining data indicative of user input via a user interface associated with a construction workflow;
determining one or more keywords associated with the user input;
determining one or more suggested content items based on the one or more keywords;
generating a predicted performance increase associated with each respective content item of the one or more suggested content items;
ranking the one or more suggested content items based at least in part on the predicted performance increase;
generating, automatically, a first constructed content item comprising at least a first suggested content item of the one or more suggested content items based at least in part on:
comparing a first predicted performance increase associated with the first suggested content item to a threshold predicted performance increase value, wherein the threshold predicted performance increase value is determined based on user input data; and
determining the first predicted performance increase is above the threshold predicted performance increase value;
augmenting a database of constructed content items to include the first constructed content item; and
performing a matching process to provide the first constructed content item for display on a second user device based at least in part on a predicted probability of user interaction with the first constructed content item.
2. The method of claim 1 , wherein the user input comprises a uniform resource locator (URL).
3. The method of claim 2 , wherein determining the one or more keywords associated with the user input comprises at least one of (i) analyzing the uniform resource locator (URL) for keywords or (ii) analyzing a website associated with the uniform resource locator (URL) for keywords.
4. The method of claim 1 , wherein generating the predicted performance increase associated with each respective suggested content item of the one or more suggested content item comprises:
generating a predicted performance increase for at least one of (i) an individual word of the respective suggested content item or (ii) a phrase of the respective suggested content item.
5. The method of claim 1 , comprising:
obtaining data indicative of a user indicating interest in a content item; and
in response to obtaining the data indicative of the user indicating interest in a first content item of the suggested content items, generating a predicted performance increase for the first content item.
6. The method of claim 1 , wherein the predicted performance increase is displayed as at least one of (i) an average predicted performance increase, (ii) a range of predicted performance increase, or (iii) a statistical significance of a predicted performance increase.
7. The method of claim 6 , wherein the predicted performance increase is a percentage.
8. The method of claim 1 , wherein the predicted performance increase is generated using a model.
9. The method of claim 8 , wherein past performance of a content item is input in the model and used to generate the predicted performance increase of each respective content item of the one or more suggested content item.
10. The method of claim 9 , comprising:
obtaining data indicative of past performance of one or more content items with one or more respective features;
comparing one or more features of a first suggested content item to one or more features of a prior used content item; and
determining, based on the comparison of the one or more features of the first suggested content item and the prior used content item and the past performance of the one or more content items, a predicted performance increase for the first suggested content item.
11. The method of claim 8 , wherein the model is a machine-learned model.
12. The method of claim 11 , wherein the model is trained using actual performance data.
13. The method of claim 1 , comprising:
obtaining data indicative of a user selection of a first suggested content item of the one or more suggested content items; and
determining a predicted performance increase of a constructed content item based at least in part of the predicted performance increase of the first suggested content item.
14. The method of claim 1 , wherein ranking the one or more suggested content items comprises:
determining a threshold predicted performance increase value;
comparing each respective predicted performance increase to the threshold predicted performance increase value;
determining that a second predicted performance increase associated with a second suggested content item is above the threshold predicted performance increase value and below the first predicted performance increase; and
providing for display the first suggested content item and the second suggested content item, wherein the first suggested content item is displayed in a more prominent position than the second suggested content item.
15. The method of claim 1 , comprising:
generating a second predicted performance increase associated with a second suggested content item;
comparing the second predicted performance increase with a threshold performance increase value;
determining that the second predicted performance increase is below the threshold predicted performance increase value; and
in response to determining that the second predicted performance increase is below the threshold predicted performance increase value, preventing display of the second suggested content item.
16. The method of claim 1 , wherein a first suggested content item of the one or more suggested content items comprises at least one of (i) a word, (ii) a phrase, (iii) an image, (iv) a video, or (v) an audio content item.
17. A system comprising:
one or more processors; and
one or more computer-readable media storing instructions that are executable to cause the one or more processors to perform operations, the operations comprising:
obtaining data indicative of user input via a user interface associated with a construction workflow;
determining one or more keywords associated with the user input;
determining one or more suggested content items based on the one or more keywords;
generating a predicted performance increase associated with each respective content item of the one or more suggested content items;
ranking the one or more suggested content items based at least in part on the predicted performance increase;
generating, automatically, a first constructed content item comprising at least a first suggested content item of the one or more suggested content items based at least in part on:
comparing a first predicted performance increase associated with the first suggested content item to a threshold predicted performance increase value, wherein the threshold predicted performance increase value is determined based on user input data; and
determining the first predicted performance increase is above the threshold predicted performance increase value;
augmenting a database of constructed content items to include the first constructed content item; and
performing a matching process to provide the first constructed content item for display on a second user device based at least in part on a predicted probability of user interaction with the first constructed content item.
18. A non-transitory computer readable medium embodied in a computer-readable storage device and comprising instructions that, when executed by a processor, cause the processor to perform operations, the operations comprising:
obtaining data indicative of user input via a user interface associated with a construction workflow;
determining one or more keywords associated with the user input;
determining one or more suggested content items based on the one or more keywords;
generating a predicted performance increase associated with each respective content item of the one or more suggested content items;
ranking the one or more suggested content items based at least in part on the predicted performance increase;
generating, automatically, a first constructed content item comprising at least a first suggested content item of the one or more suggested content items based at least in part on:
comparing a first predicted performance increase associated with the first suggested content item to a threshold predicted performance increase value, wherein the threshold predicted performance increase value is determined based on user input data; and
determining the first predicted performance increase is above the threshold predicted performance increase value;
augmenting a database of constructed content items to include the first constructed content item; and
performing a matching process to provide the first constructed content item for display on a second user device based at least in part on a predicted probability of user interaction with the first constructed content item.
19. The non-transitory computer readable medium of claim 18 , wherein the predicted performance increase is generated using a model, and wherein past performance of a content item is input in a model and used to generate the predicted performance increase of each respective content item of the one or more suggested content items.