📘 This article provides information on how ActionBoard priorities are determined. ActionBoard is part of Botify's Intelligence Suite, available with all Botify plans.
Overview
Actions in ActionBoard are prioritized to help you quickly determine which actions to take first, based on their estimated impact. Botify uses both machine learning and static criteria to evaluate priorities. The action is prioritized using machine learning criteria when the model confidence score is high, simulating how many additional pages would be crawled or active if the issue is fixed. When the confidence score is not high enough, or machine learning is unavailable for the action, the priority is derived from the static criteria.
Both logs and visits data (from logs or analytics GA, Adobe, Piano) are required for machine learning priorities in ActionBoard. If you do not have both logs and visits data integrated with Botify, you will only see priorities based on the static critiera in ActionBoard.
Filtered Actions
Priorities are calculated based on your entire site by default. When you filter ActionBoard by specific segments or breadcrumbs, the priorities and machine learning impact are recalculated based on those selections. If you fix the issues specific to the selected segments or breadcrumbs, this reveals the expected impact on crawled and active pages.
Machine Learning Priorities
ActionBoard currently computes the impact of 41 actions by machine learning, evaluating for crawled and active pages.
When the model confidence score is high enough, the machine learning criteria will be displayed in ActionBoard instead of the static criteria. Machine learning scores are calculated separately for each machine learning criterion, and priorities are displayed if one of the following conditions are met:
The impact for both crawled pages and active pages is high
The model confidence score in the simulations for crawled pages or active pages is high
This means that some crawls for an ActionBoard action may show the machine learning impact while others may only show the static priority, and vice versa. Machine learning was introduced to ActionBoard as of the 4th November 2022 release, and impact scores were not computed for crawls before this release.
Note: Although we include only the machine learning criteria with the highest confidence scores in ActionBoard, they are based on simulations; therefore, you should not expect them to be perfect, as with any machine learning.
Machine Learning Priority Thresholds
The following are the thresholds used to determine the machine learning priorities:
High: +10% estimate of additional crawled/active pages opportunity
Medium: between +5% and +10% estimate of additional crawled/active pages opportunity
Low: lower than +5% estimate of additional crawled/active pages opportunity
Current Machine Learning Actions
The following are the actions that are currently evaluated by machine learning.
Ensure That Description Is Not Duplicated
Set Meta Description on Pages
Increase Meta Description Length
Ensure that H1 Is not duplicated
Set H1 Heading on Pages
Ensure that Title is not duplicated
Set Title Tag on Pages
Increase Content Length in Title Tag
Address Thin Content on Indexable Pages
Improve Load Time on Non-Canonical Pages
Improve Load Time on Indexable Pages
Improve Load Time on 4xx Pages
Increase Inlinks on Indexable Pages
Bring Page higher in Website Structure
Fix Links in Content pointing to 3xx Pages
Fix Links in Content pointing to 4xx Pages
Fix Links in Content pointing to 5xx Pages
Fix Links in Content pointing to Pages with bad Canonical
Fix Links in Content pointing to Pages with noindex Tag
Fix Links in Template pointing to 3xx Pages
Fix Links in Template pointing to 4xx Pages
Fix Links in Template pointing to 5xx Pages
Fix Links in Template pointing to Pages with bad Canonical
Fix Links in Template pointing to Pages with noindex Tag
Verify Accuracy of Mobile Pages with less than 50% Similarity
Remove Canonical tags from Canonical Pages with less than 50% Similarity
Verify Accuracy of Mobile Pages with less than 50% Similarity
Resolve duplicate or highly similar Content
Diversify Anchor Text
Improve Load Time to < 2 Seconds on 3xx Page
Improve Load Time to < 2 Seconds on 4xx Page
Improve Load Time to < 2 Seconds on 5xx Pages
Improve Load Time to < 2 Seconds on Non-Canonical Pages
Improve Load Time on 5xx Pages
Improve Load Time on 3xx Pages
Improve Load Time to < 1 Second on noindex Pages
Remove 3xx Pages from XML Sitemaps
Remove 4xx Pages from XML Sitemaps
Remove 5xx Pages from XML Sitemaps
Investigate Links to Pages blocked by robots.txt
Resolve Hreflang Errors
Static Criteria Priorities
When the machine learning priorities cannot be used (because ML is unavailable for the action or the confidence score is not high enough), the static criteria are displayed instead. The static criteria are volume of URLs, visits, crawl budget, and PageRank. Some actions may only have one or a subset of these criteria that apply.
Static Criteria Priority Thresholds
The following are the thresholds used to determine the static criteria priorities:
PageRank
High: The proportion of PageRank consumed by impacted pages strictly greater than 10%.
Middle: The proportion of PageRank consumed by the impacted pages strictly higher than 5%.
Low: The proportion of PageRank consumed by the impacted pages lower than 5%.
Volume
High: The proportion of impacted URLs strictly higher than 10%.
Middle: The proportion of impacted URLs strictly higher than 5%.
Low: The proportion of impacted URLs lower than 5%.
Visits
High: The difference between the percentage of inactive pages impacted and the percentage of active pages impacted is strictly greater than 10%
Middle: The difference between the percentage of inactive pages impacted and the percentage of active pages impacted is strictly greater than 5%.
Low: The difference between the percentage of inactive pages impacted and the percentage of active pages impacted is less than 5%
Crawl
High: The proportion of crawled URLs impacted strictly greater than 10%.
Middle: The proportion of crawled URLs impacted strictly above 5%.
Low: The proportion of crawled URLs impacted less than 5%.
About “None" Priority Actions
Actions are identified with “None” priority for any of the following conditions:
The issue is not detected anymore, but it was detected in the compared crawl.
The action requires visits data, and your visits data is not yet integrated with Botify. Please contact Support to integrate your visits data with Botify.
The action only has Crawl criterion, and the issue does not affect pages crawled by Google.
The action has Visits & Crawl criteria, and the issue does not affect visits probability or URLs crawled by Google.
Read next: