In this article
Share of Voice (SOV) is often treated as a report card demonstrating whether a PR campaign generated more media coverage and gained visibility over its competition. Using SOV to evaluate campaign effectiveness is certainly one use, but it is not the only one. Share of Voice reports can be diagnostic tools that reveal competitive opportunities, campaign anomalies, seasonal patterns, and messaging gaps.
This article explains how to find and use this hidden information that is produced in the more practical SOV reports typically created by PR teams. Such reports are not intended to replace Business Intelligence analysis, but they can become significantly more strategic with the right setup and interpretation.
First, let’s dissect the elements of Share of Voice and how to set up the reporting so it can be a valued and trusted measure of campaign performance.
Share of Voice compares brand visibility in the media. Basic Share of Voice (SOV) compares the visibility of a product or brand in the media relative to a defined set of competitors during a specific period of time. It is usually produced in association with a media database that can search current and historical presence in articles and media coverage based on search strings. SOV for PR purposes usually counts the number of articles and media coverage (broadcast, radio, newsletters, and podcasts). Sometimes those reports include AVE (advertising value equivalency), sentiment, and audience numbers.
Advanced Share of Voice reports can add complexity by separating product lines, and assessing message pull-through. It is also possible to ascribe a weighting system for high-value media outlets based on their audience and domain authority. It can also segment coverage by category, such as trade publications, national media, podcasts, and newsletters.
As with any data project, protocol and standardization matter so consider these elements when setting up your Share of Voice reporting system.
Set the list of competitors being monitored. Surface every competitor that needs to be monitored early and include them in the baseline report. You’ll want to maintain this as the baseline for every subsequent report. It is possible to add competitors later but it will skew the percentages. That is fine, but make note of it and if decisions are being made about ROI when a new competitor was introduced mid-campaign, run that ROI report with and without the new competitor so that a fairer comparison can be made.
Running an additional SOV report, though, does not fully account for the new competitor. If a new brand or product entered the scene mid-course and it has drawn significant media attention, that will skew the results for both your brand and the other established competitors.
A note about competitors: There are different kinds of competitors, but it is not a good idea to mix them into the same report. Competitors are not always products. Depending on the objective, they may be alternative solutions, substitute approaches, aspirational brands, or organizations competing for the same conversation. There could be aspirational competitors or competitors in a category. Such depth may fall under BI’s purview, but have these conversations early with your client so that the baseline SOV report is measuring the right competitor.
Refine the Search Parameters Early. Setting the search parameters correctly at the beginning is essential because this is the baseline on which all other success will be measured. The numbers here get locked in (with a few exceptions) so be prepared to take several run-throughs to find the search strings that accurately assess your brand’s visibility and each competitor’s. It is likely that each company in the report has additional search parameters to filter out noise and confounding factors.
It might be helpful to run a few searches on the space alone to see if the terms are surfacing the intended articles. For example, running a search for articles on “middle school algebra programs” or “learning management systems for higher education” can surface noise that needs to be corrected with each individual brand. The category being monitored (for example, “middle school algebra programs”) remains constant, while each brand requires its own qualifying search terms to reduce noise.
Companies with common-language product names like Prodigy or SpringBoard might need many more qualifying search terms than a product with a unique name. Run these search terms several times in your media database until you feel confident you are capturing everything without undue noise.
HINT: Use ChatGPT or Claude to help refine search strings. Load your search string directly into the prompt window and ask it to suggest additional terms if you discover noise in the database. The AI might return Boolean operators you didn’t know existed.
De-dupe and account for syndication. Syndicated articles are those that are published in multiple outlets. Some media databases have a simple toggle to remove syndicates but they are not always 100 percent accurate. Manual review is needed and it can be laborious. However, it is essential to de-dupe (for every brand). Regularly checking your trackers and de-duping once a week is well justified.
Be consistent in deciding to count press releases (or not). Choosing to count press releases can be subjective, but what matters is consistency. Just as for syndicating articles, your choice for press releases must be applied to every competitor, which could mean evaluating thousands of articles. We include press releases in our SOV reports because of the intelligence that information provides, but we exclude anything that is a verbatim pickup citing the wire as the source, as well as the syndication feeds.
Once the baseline is established, avoid changing search strings, but if they do, here’s how to deal with it. If the search parameters need to change, immediately run a new baseline report using the original search time period. Your aim is to recreate the baseline report with the new search string. Then apply a file naming convention to keep clear on which system you are working. Make similar notations on your reports visually. Keep the discarded SOV reports with the old search parameters because you may want to refer to them later. Save these in a PDF with downloaded spreadsheets because if your database period drops off old data, you’ll not be able to recreate that report without moving to a new system or paying to recover a longer look-back period.
Document the reporting protocol. Document the search string protocol in such detail that someone new to the project could run the exact same report for a new time period. Better yet, add an abbreviated version of that including the search strings to each SOV report.
Uncovering Hidden Insights in Share of Voice Reports:
When the baseline SOV report is solid and the tracking tools are operating at their best efficiency, it’s time to uncover the hidden insights and consider how they might work into your public relations strategy.
Pie Charts show which brand has the largest share of voice, but bar charts help explain why.
The Pie Chart showing relative strength of each brand helps give a visual sense of performance over a certain time period. A Bar Chart produces intelligence that can stand by itself and it prompts a deeper investigation into what has been occurring over a specified time period.
The first step is to look at the bar chart for outliers, which are any portions of the charts that are starkly different from the typical pattern. In the fictitious sample below, notice that in August 2025, the Green company spiked in coverage. That should inspire looking more closely at that company’s coverage to see what happened. Did they issue a press release? Did they publish a data report? Is there a syndicated story that is being overcounted? If so, fix that and rerun the report.
Other reasons for a spike could be that the competitor held a contest or a promotional event. If so, then create a campaign to pull back attention in the same time period. Or, run a better event before your competitor’s. By the same token, the competitor may have been subject to a period of negative media coverage. If you hadn’t recognized that this was happening before, now is the time to investigate.
In our fictitious sample, there is also an interesting trend of increased coverage for the Orange company. Unlike the August spike for Green, Orange is showing a deliberate effort to increase visibility. They may have engaged in a thought leadership campaign or in increased media relations following a product launch. These are the data periods to study more closely to see if a competitor’s messaging has changed and is subsequently being well received. If market share is also growing along with Share of Voice, investigate why more closely.
Noticing these trends and outliers is the first step to turning insights into more effective strategy.
When you see an outlier, consider these questions:
- Was there a press release or major announcement?
- Did a search parameter fail?
- Was the coverage earned or syndicated?
- Did sentiment change?
- Which publications drove the increase?
- Is this seasonal?
- Is it repeatable?
- Does it require a response?
Finally, there are more metrics to consider like AVE, audience and sentiment.
Above and beyond the number of articles which we used in these examples, AVE, audience and sentiment may also be helpful. AVE should not be treated as a precise measure of business value, however, it can be helpful if no other agreed-upon metric has been established. Many executives still ask for AVE, so consider including it as a secondary benchmark, rather than the primary proof of performance.
Running Share of Voice on audience numbers and sentiment may also be of interest. Lock these into your baseline report and discuss how they may be helpful to those responsible for the PR budget.
Sentiment, though, comes with its own set of issues. It is a useful measure especially when there is any controversy around a brand but sentiment is subjective and therefore unreliable with automated sentiment analysis. Getting an accurate measure of positive, neutral, or negative requires that a human read (or at least skim) every article in the report to manually adjust the sentiment ratings that are incorrect. It is time-consuming, but may be necessary. Try to ascertain if sentiment will be a critical metric before getting too far into the project so that you can allocate sufficient hours to the account and establish a regular review schedule.
Share of Voice in its various iterations is a valuable measure of the progress and effectiveness of a public relations campaign. Don’t submit to the temptation to treat it like a monthly scorecard, though. Did we win the month? Are we beating our competitor this month since we weren’t last month? Treating it this way misses its greatest value. The most useful SOV reports raise questions and point out instances that need further investigation. Used this way, Share of Voice becomes less of a scorecard and more of an investigative tool—one that helps PR teams understand competitors, the media landscape, and where the next opportunity may exist.
This was originally published in PR in EdTech on LinkedIn on July 9, 2026.
Share This Story, Choose Your Platform!
Subscribe to receive tips
There are two ways to subscribe to our PR in EdTech newsletter:
- Subscribe on LinkedIn and see extra content like recent articles on Substacks, and what to do after an article lands in a publication.
- Subscribe via email to receive news earlier and with few more details not shared widely on the LinkedIn newsletter.