Is there enough social video about your brand? How to check before you buy a tool

The question most procurement conversations fail is not "how many posts?" It is "what decisions do we need to make, what happens if we miss one, and how much relevant content is actually there to monitor?" Volume is one of three inputs. On its own, it is a weak guide.

The method below runs the volume check honestly, in a bounded 30-minute first pass, and then feeds the volume into a buying rule that includes risk and decision frequency alongside it.

The buying principle

Match the monitoring format to the decisions you need to make, the consequences of missing a post, and the amount of relevant content available.

A brand with ten consequential videos a month (reputational risk, product-safety signals, regulatory exposure) may need alerts more urgently than a brand with thousands of routine mentions and no video that threatens the business. Format follows need, not volume.

Which brand and category terms to count

List three sets of terms. The brand itself: name, common misspellings, product names, handles. The category: what buyers call the thing you sell, not what your marketing calls it. And the competitors you would compare against.

If your brand name is also an everyday word, add context terms so the count is not inflated by unrelated posts. The piece on a brand name that is also a common word covers how to disambiguate without over-pruning. Write the terms down. The same list goes to every vendor so the counts are comparable.

A fair time window

Count at least three months, and twelve if your category is seasonal or launch-driven. One product drop or one viral post can make a single month look busy. Note the dates and time zone, and ask each vendor how far back their history goes for your plan.

Compare like with like: a 90-day count from one tool against a 30-day count from another tells you nothing.

Splitting the count: four exclusive buckets

Split every count into four buckets, assigned in order so each post lands in exactly one.

1. Own accounts. Posts from the brand's own handles. Everything else moves to step 2.

2. Known paid. Of the remaining posts, those from creators on your contract or brief list. Everything else moves to step 3.

3. Disclosed commercial. Of what remains, posts carrying a platform disclosure such as TikTok's "Paid partnership" or "Commercial content" label, or the equivalent on Instagram (paid partnership label). Everything else moves to step 4.

4. Unknown commercial status. Everything that did not land in a bucket above.

The sequential rule matters: a creator you commissioned who also carries a platform label counts as known paid, not disclosed commercial, because step 2 fires first. The absence of a disclosure is not evidence of its absence; creators may underdisclose and platforms do not catch everything. The fourth bucket is honest about what you cannot determine from the content alone.

Each bucket tells you a different thing. Own accounts shows your activity. Known paid shows what you bought. Disclosed commercial captures paid work by creators you did not commission (or caught by platform rules). Unknown status is the harder read: usually the market, sometimes undisclosed paid.

Posts vs. views

Forty posts a month can matter more than four hundred if a handful reach millions. Report posts and views on separate lines, and note that attention is usually highly skewed.

Name the metric each platform reports. On Meta, reel plays are now called "Views", which changes what the same number means in a report versus in the app.

For scale, dig's Skincare 2025 report covered 51,583 posts over eight months, and the Black November report covered 69,403 posts from 31 October to 1 December 2025. See the skincare report for the method. These are whole categories across many brands, a reference point rather than a target for one brand.

Speech-only mentions: a bounded check

In video, people often say a brand name without typing it in the post description or adding a hashtag. A tool finds those mentions only if its pipeline transcribes and indexes speech.

Two questions to ask every vendor. One: does your count include posts where the brand is mentioned only in speech? Two: can you show me which posts in my test window were speech-only, so I can check?

Then run the bounded discovery test: find 20 posts in your category by hand (platform search, your own feeds, team members, your inbox), note which say the brand only out loud, and before supplying any links to the vendor, check whether the tool's own results already include those posts. Supplying the links tests how the tool reads a post. Checking whether the tool found the post independently tests discovery, which is the question that matters here. The result is a bounded check on those 20 posts; it is not a market-wide finding and not a universal rate for your category.

Worked example: setting a buying rule, not a volume threshold

Illustrative only, not a dig benchmark.

Marlow Ceramics, a fictional cookware brand, counts 1 August to 31 October.

Category term "ceramic cookware": 9,800 posts in the same window, about 3,267 a month.

Marlow's rule, written before counting (not based on the number): the comms team must respond within one business day to a potentially reputational post, and within one week to a product-feedback post. Potentially reputational posts in the last year averaged about 4 a month. Product-feedback posts, roughly 40 a month.

With ~128 unknown-status posts a month to triage, a quarterly brand report misses the reputational response window. A monthly report misses it too. The buying decision is not a threshold on monthly volume; it is that Marlow needs alerts within a day on potentially reputational posts and a weekly digest for the rest. Always-on brand monitoring with risk-based alert thresholds is the match; category-level tracking runs alongside to catch rising themes.

If Marlow had been a brand with 50 posts a month and no reputational history, the same rule would have led to a monthly category report with ad-hoc checks. The volume did not change the rule; the decisions it needed to make did.

How long the volume check takes

A first count on a brand and a category takes roughly 30 minutes in a tool that lets you run ad-hoc queries. Validation (relevance of the hits, commercial-status assignment, discovery test on the 20 hand-found posts) is a separate pass. As a planning allowance, budget about half a day per vendor for validation; the actual time varies with the vendor, the category and how clean the first count was.

ask-dig answers a quick brand or category question on TikTok, which gets you a first count without a procurement call. Free tiers of other tools can also give a rough first count; check what each one includes. Always-on monitoring across networks is dig Enterprise.

Once you have the count, set up tracking on the terms you just tested. The piece on how to track and measure brand mentions in social video covers what comes next.

Pro tip

Match the monitoring format to the decisions you need to make and the consequences of missing a post. Volume is one of three inputs, not the input.

Key takeaways

• The buying rule is decisions × risk × volume, not volume alone.
• Count brand and category terms over at least three months; compare like windows across vendors.
• Split into four buckets: own accounts, known paid, disclosed commercial, unknown status. Absence of disclosure is not evidence of absence.
• Report posts and views on separate lines.
• A 20-post speech check tests the tool on those 20 posts, not the whole category.
• A 30-minute first count is the start; validation takes another half-day before you quote the number.

Frequently asked questions

How many social media mentions is enough for a monitoring tool?

There is no universal number. Decide the decisions you need to make, the consequences of missing a post, and then compare against your monthly count. A low-volume brand with reputational risk may need always-on alerts; a high-volume brand with no risk-sensitive content may be fine with a monthly digest.

Should brand-owned posts count as mentions?

Count them in their own bucket. Your own posts show your activity, not the market's. Paid, disclosed commercial, and unknown-status posts each tell you something different, and sweeping them together into "mentions" obscures what the number means.

Can a mentions tool find brand names that are only spoken in a video?

Only if it transcribes and indexes speech. Ask the vendor directly, then run the bounded test: pick 20 organic videos you found by hand, note which say the brand only out loud, and check whether the tool's results include them. The result tells you how the tool performed on those 20 posts.