The question this article answers
A memory dashboard can become impressive very quickly. Four measures, several colours and a trend line later, nobody is sure which buyers were counted. Mental availability metrics are only useful when their denominators are clear. The practical task is to understand how many category buyers connect a brand with buying situations, how many situations they connect, and how much competing memory is present. The calculation must preserve those different questions rather than compress them into one attractive score.
Build the association matrix first
Begin with researched buying cues and a defined competitive brand set. Ask eligible category buyers which brands they associate with each cue, allowing an appropriate none option and multiple selections. Each respondent-brand-cue connection is one recorded link. A person selecting the same brand under three different cues contributes three links, not three people.
Retain the underlying respondent-level data. A totals-only export may support mental market share but leave you unable to identify the people with at least one link, which you need for the other calculations. If respondents are weighted, apply a consistent analysis convention and document it. Keep the raw and weighted outputs distinguishable rather than mixing them in the same example. Read our guide to brand salience for the commercial question behind the calculations.
Four measures and their denominators
Mental market share is the brand’s links divided by all measured brands’ links across the full eligible sample. Mental penetration is the number of people with at least one link to that brand divided by the eligible sample. Network size is the brand’s links divided by the number of people with at least one link to it.1 The units differ: shares are percentages; network size is a count of links per person in that subset.
Share of mind restricts the competitive comparison to people who have at least one link to the focal brand. Within that subset, divide the focal brand’s links by all brands’ links. Romaniuk explains the reason for this restriction: it examines competing associations among people who already have the brand in memory.2 Report whether you use this pooled-link calculation; do not silently substitute an average of individual percentages. Distinguish these tasks from recognition in brand awareness vs mental availability.
The same percentage can hide different patterns
A brand can add associations among a small familiar group while failing to appear among more category buyers. Another can reach more people but acquire only one narrow association in each person’s memory. Those patterns suggest different next questions. The first calls attention to audience reach; the second calls attention to the situations portrayed and the offer’s breadth.
Mental availability metrics should be read together, with actual buyer and non-buyer results where recruitment permits. A change in one metric does not dictate a creative answer. Examine sample changes, competitive activity, category participation and whether the business can serve additional situations before deciding what production or media should do next. Research category entry points before building the association matrix.
Keep the measurement instrument stable
Adding more cues creates more chances to produce links. Adding competitors changes the denominator for competitive shares. Either can alter results even if buyers’ memories have not changed. A dashboard without a method note can therefore imply progress where only the questionnaire expanded.
Freeze the core cue set, brand set, eligibility and fieldwork procedure for a comparison series. Where a genuine market change requires an update, use a documented bridging design or start a clearly marked new baseline. Keep subgroup sizes visible. A striking result among fifteen buyers should remain an exploratory observation, not become the agency’s confident recommendation for the entire market.
A manual check before the dashboard
Take three fictional respondents and a few cues, then write out every selection in a small test file. Include somebody with no focal-brand links, somebody with one and somebody with several. Add competitor selections for the people who have focal-brand links and for the person who has none. This deliberately small dataset makes the subset differences visible.
Calculate all four measures by hand before reviewing the script or platform output. Check that the zero-link respondent remains in mental penetration’s denominator but not network size’s. Check that competitor links from that person appear in full-sample mental market share but not in the focal brand’s restricted share of mind. Those checks catch errors that otherwise survive because the final percentages look plausible.
Next, test a missing answer and a none selection. They are not automatically the same thing. A respondent who deliberately selects none completed a task; someone who skipped it may have an incomplete record. The exclusion or retention rule needs to be agreed and applied consistently, particularly when comparing waves.
Preserve this synthetic test alongside the reporting specification. When a research platform, analyst or questionnaire changes, rerun it. A calculation that was correct last year can break after a new export format or weighting step. This simple check gives the client a practical way to verify the dashboard without needing to inspect every production data row.
Check the denominator before presenting the dashboard
Share the questionnaire, brand set and calculation definitions. We will discuss a clear memory-measurement brief and the decisions it should support.
Worked example using fictional numbers
Illustrative example only: 400 eligible respondents create 2,400 links across the measured brands and cues. Fictional brand Solen receives 480 links from 160 people. MMS is 20%, MPen is 40%, and NS is three. Among those 160 people, all brands receive 1,200 links in total, so Solen’s SoM is 40%. The last denominator includes competitor links only from that restricted group. Dividing 480 by 2,400 again would merely repeat MMS. The sample and figures are invented to show the arithmetic, not to suggest desirable targets.
Putting the evidence to work
Write a calculation specification before fieldwork. Include the category definition, cue list, brand set, unit of analysis and treatment of incomplete responses. Test the analysis with a tiny synthetic dataset in which each answer can be counted by hand. This catches denominator errors before a polished report hides them.\ \ A Quiamo assignment could combine the study specification, an editorial explanation of findings and a creative response brief. Commission specialist research support where the sampling or weighting needs it. The useful handover includes the full definitions, comparable baseline and decisions the evidence supports. It should also state the decisions it cannot support, particularly claims about causal sales impact. Explore how Quiamo approaches brand strategy and communication before defining the scope.
Questions people ask
Is share of mind the same as mental market share?
No under the convention used here. Mental market share uses links from the full eligible sample. Share of mind uses all-brand links only among people who have at least one association with the focal brand. Because the subset changes by focal brand, share-of-mind percentages across brands should not be added as if they describe one shared market total.
Can network size be averaged across everyone?
You can calculate such an average, but it is a different statistic. Network size as defined here excludes people with zero focal-brand links from the denominator. Label an all-sample average explicitly if you report it. Otherwise readers may think breadth among familiar people improved when the actual change was simply more people having any association.
Do these metrics replace an effectiveness experiment?
No. They describe measured memory structures. To argue that a campaign caused a change, the study needs a suitable causal comparison. To argue it caused sales, measure the relevant business outcome with an appropriate design. A dashboard can inform a hypothesis without proving every link between exposure, memory and purchase.
Where to go next
Before adding mental availability metrics to a board report, make the calculation readable enough to challenge. Bring Quiamo the questionnaire, data definitions and commercial decision at stake. We can discuss a measurement brief that keeps the arithmetic honest and gives creative teams a useful response to the findings. Contact Quiamo to discuss the brief.
Make the memory dashboard auditable
If the dashboard already exists, bring its definitions and a synthetic calculation example. We can discuss the checks needed before its figures enter a management report.
Sources
- Category entry points and memory metric definitionsMMS, MPen and NS denominators.
- Jenni Romaniuk interview on brand trackingOriginal interview explaining restricted share-of-mind subset.
Worked examples on this page are fictional and labelled. Search phrases and figures from external sources are cited; nothing here describes a Quiamo client result.
