Both Perplexity and Copilot retrieval systems are experiencing outages. The team is investigating and will follow up with more details.

As a follow-up to our status update last night, we wanted to send an additional update talking about next steps and the new reality of SERP data after this latest round of Google actions.

We resolved / closed the previous incident because that had been open for more than a month and was initially scoped around the total blockage that Google had initiated.

Since the last week of September, however, we have been providing daily data for all accounts, to varying degrees of success depending on market (Japan was the most difficult).

The main difference was that we have been providing it for SERP page 1 in most cases except for Desktop (outside of Japan), where we have been continuing to provide 5 pages.

The technical difficulties are largely dealt with, although there are some ongoing situations we continue to monitor, such as the spurious / poisoned SERP data, empty AIOs for about 30% of the cases, and so forth. We will continue to work on these.

The main factor now that has changed, for the whole industry is cost relative to overall throughput. It's not about cost only and not about throughput only, but the two factors combined. This means that fetching teams are focused on optimizing for the most throughput for the most accounts at the best possible cost, in order to continue offering this data without impacting customer costs.

We will be reaching out to customers to explain things in more detail but the main effect of all of this is that we are planning to reduce the number of pages in the SERPs we fetch, in order to keep overall fetching throughput and success rates high.

We have spoken with a number of other leaders in the space and we believe that this is both the direction that the industry is heading and also where the most actionable data lives. The 5 page depth was more or less an artifact of the previous 10 page depth that everybody got for free.

This also means that our Top 20 data, both via API and in Search Intelligence (BigQuery) will be limited to the first page, so it's more like Top 10 data until this is resolved.

Further details will be provided in calls with clients directly, but please feel free to reach out with any questions.

The primary purpose of this incident report is to make it clear that we are still fetching page 1 only for most sites and accounts but are taking steps to go to 2 pages as quickly as possible.

We also noted this in our previous update but felt that it warranted an incident report of its own.

We've reached a point of stability where we have been reliably fetching the 1st page of the SERP daily and are working now to expand this to the 2nd page. We're still evaluating options for SERP page depth but feedback from clients so far has centered on the importance of the first two pages, so we're focused on that.

We are moving the status of this system back into fully operational as a result but we are continuing to work around the clock to further boost data availability as deep as we're able to go.

The challenges at this point aren't technical, we're able to get the data. It's a matter of costs, which have increased dramatically since last month. We're working to do whatever we can to reduce these costs in order to provide the most actionable data possible. We thank you for your continued understanding.

Over the last week, we have achieved a level of stability, although we are continuing to work to improve things.

Desktop is in a better place than mobile. Japan is the most challenging at the moment.

We are currently fetching page 1 for all sites / all keywords at the moment. For desktop, outside of Japan, we are currently able to fetch the top 5 pages about once every 5-6 days. We have implemented our sampling regime to provide coverage at the sub-page level, meaning that the core ranking data shouldn't have major gaps on a day to day basis.

AI Overviews are a challenge as well, and we are working to improve coverage here too. As with the rest of the situation, mobile is more of a challenge than desktop is.

We have also started to see indications that Google is returning spurious results in the SERPs. This is not yet fully widespread but it is also on our radar and clients should be aware that this is happening.

As always, we appreciate your patience and are happy to schedule a call with any customer to walk through any questions.

We are continuing to work on restoring fetching. We have made some progress on Desktop, Mobile continues to be a challenge. Data processing is still not fully restored, we will continue to provide updates.

We are continuing to work to resolve the issues related to this outage, but do not yet have a concrete timeline to share.

We have been able to confirm with other providers in the space that this is having a widespread impact. A number of providers (including DemandSphere) are collaborating in order to collectively restore access as quickly as possible.

We appreciate your patience. Please email support@demandsphere.com with any questions.

In addition to the empty AIOs, we are anticipating a delay in data for September 10th-13th.

We are seeing a large volume of empty AI Overviews. We are investigating the situation and will provide updates as we have them.

We have deployed a countermeasure to these changes and are testing it in production. So far, we are seeing positive results and expect this to continue. There may be some slight delays and volatility for the next day or two. Please contact your account manager with any questions.

Google has rolled out large-scale masking of URLs on the SERPs, including in AI surfaces, which will make it challenging to accurately generate ranking data. We are working through a number of countermeasures in order to restore full functionality, but please be aware that ranking / SERP data will experience volatility over the next few days.

We will provide further updates as we have new developments.

The PostgreSQL database cluster repairs have been completed successfully, and all services should now be back to fully operational. Our team has verified that systems are running normally and performance has been restored. If you continue to experience any issues, please don't hesitate to reach out to our support team. Thank you for your patience and understanding during the maintenance window. We apologize once again for any inconvenience caused.

We are currently performing emergency repairs on our PostgreSQL database cluster. This maintenance is expected to take several hours to complete, during which you may experience degraded performance or intermittent service disruptions. Our engineering team is actively working to restore full functionality as quickly as possible, and we will provide further updates as the situation progresses. We sincerely apologize for any inconvenience this may cause and appreciate your patience and understanding while we resolve this issue.

The secondary vacuuming process has completed successfully, so we are running full speed to catch up on all data deliverables as quickly as possible.

We will keep this incident open in Monitoring status until things are caught up.

We had caught up on most data until 10/14/2025.

A secondary process kicked in for our primary Postgres cluster needing vacuuming, which occurred due to a buildup from the events of the last few days.

This is an unavoidable process in order to avoid extended application downtime so we are experiencing a new delay while this process completes.

It is currently 80% complete, but it is not clear yet when it will finish. The vacuuming process has already been running for 24 hours.

Data is currently delayed for 10/15/2025 and we expect further delays for 10/16/2025.

Based on the backlog as a result of this, we will be working through this weekend to back fill data.

There are no issues with fetching data, it is a reporting and UI delay only.

We have completed reporting data for 10/12 and are focused on completing everything for 10/13 next. Our goal is to prioritize the most recent data first as this is what reflects client's latest website visibility.

We are aware that there are still gaps for some sites on 10/08-10/11. We will continue to backfill these as the system catches up.

Competitor Discovery and GA4 data are missing for most of the affected date range (10/08-10/13). We will backfill these once the core reporting subsystems are populating normally.

As an update, we have identified the issue (which we will explain below) that caused the delays to occur.

We have been fetching since this incident began without incident so we do not anticipate any missing data, we are working to resolve the delay in the display and availability of this data.

The following dates are still showing delayed or missing data. Once we complete the data recovery process, these dates will have complete data, there is simply a delay in their processing.

• October 8th
• October 9th
• October 10th
• October 11th
• October 12th

On October 8th, our primary distributed job queue manager cluster began to experience errors. We made several attempts over the next few days to repair it but it became clear that we would need to perform a larger upgrade. Specifically, we added additional redundancy and created separate queue manager clusters per subsystem (divided into fetching, parsing, reporting, and ui).

This, combined with further optimization of our parsing process, has improved overall stability and performance greatly. Unfortunately, during this time, the reporting and display of this data fell behind, leading to the delays currently showing in the app. We are working to back fill and have the entirety of this data caught up as quickly as possible. We will continue to provide updates until it is resolved.

Please reach out to your Account Manager with any questions or if you need data for specific dates, in the meantime.

We apologize for this inconvenience.

The DemandSphere Team

Due to an issue with our worker processing infrastructure, there will be a partial delay on data for October 8th, and a complete delay on data for October 9th-10th.

We are making the necessary upgrades and improvements to resolve the issue and we will provide further updates as we get closer to resolution.

We apologize for the inconvenience.

There is a delay in our Discovered Competitor dataset, about 3-4 days, depending on the site in question.

There is not an issue in data acquisition, simply on displaying the data in the UI. We will provide an update once this is resolved.

Resolved

We have identified an issue with reduced citations occurring on Google AI Overviews, for mobile devices only.

We have improved the situation by 25% in the last 24 hours but are still working to improve results further.

We apologize for the inconvenience.

We are encountering issues with our PostgreSQL cluster. Our team is working on fixing this issue.

We sincerely apologize for any inconvenience and appreciate your patience.

We’re currently experiencing delays with our 03/19 Ranking Reports, which also affect related screens like Serp Feature Explorer, Potential Traffic, and Average Rankings. The root cause is an ongoing issue with our OVH object storage. We’re continuously monitoring the situation and are in contact with the provider. We sincerely apologize for any inconvenience and appreciate your patience.

We found and fixed cause of the issue.

We're experiencing issues with our database cluster. Our team is working on it.

We encountered an incident with our authentication server which caused 54 minutes of downtime today. We have resolved the issue causing the downtime and are performing root cause analysis to mitigate further incidents. We apologize sincerely for the inconvenience.