What Belongs on a Competitor Dashboard, and What Does Not
One Tile Per Decision, Not One Tile Per Available Metric
Short answer
A competitor dashboard should hold one tile per recurring decision, not one tile per available metric. Six to ten tiles covering changes, posting cadence, ad activity, pricing and review movement, plus your own numbers on the same fields, is enough for a small team to run a weekly review.
1. What belongs on a competitor dashboard?
Eight tiles cover almost every recurring competitive decision a small marketing team makes: changes, cadence, ad activity, ad messaging, pricing and page changes, review movement, share of the tracked set, and your own numbers on the same fields.
- Competitor dashboard
- A single screen holding the competitive measurements a team reads on a fixed rhythm, where each tile exists to inform a decision that recurs rather than to display an available number.
| Tile | Question it answers | Refresh rate |
|---|---|---|
| Changes since the last review | What did any tracked competitor do that I have not seen yet? | Collect daily, read weekly |
| Posting cadence per competitor per channel | Who is publishing more or less than they normally do? | Weekly |
| Active ad count per competitor | Who is putting more weight behind paid right now? | Weekly |
| New ad creative and new claims | What is each competitor paying to say this month? | Weekly |
| Pricing and key page changes | Has anyone changed what they charge, promise or lead with? | Collect daily, alert on change |
| Review volume and rating direction | Is anyone accumulating complaints, or running a review campaign? | Monthly |
| Share of the tracked set on one metric | Are we gaining or losing ground inside the set we chose? | Monthly |
| Our own numbers on the same fields | How do we look when measured exactly the way we measure them? | Weekly |
The eight tiles map to eight questions someone on the team asks out loud most months. That is the test for adding a ninth: name the question and name who asks it. A tile that cannot pass the test is a chart, not a dashboard component.
2. What should be left off a competitor dashboard?
Anything modelled rather than observed, anything with no owner, and anything already delivered as an alert. Each of those either invites a wrong decision or trains the team to skim past the whole screen.
Six things that make a dashboard worse by being on it:
- Estimated ad spend and estimated website traffic. Both are modelled from panels rather than observed, and on a dashboard they lose the caveat within a week and get quoted as fact.
- Competitor follower counts as a headline number. They move slowly, they are trivially inflated, and they change no decision anyone makes.
- A single sentiment score. Compressing review and comment text into one number destroys the only useful part, which is what people are complaining about.
- Anything already covered by an alert. If a pricing change fires an alert, the dashboard tile duplicating it just adds a second place to look.
- Totals summed across channels. Cross-channel sums of unlike units are dominated by the largest raw counts and hide which channel moved.
- Any metric with no named owner. An unowned tile is never investigated when it moves, so it eventually gets ignored when it moves for a real reason.
3. How do I build the dashboard, in what order?
Start from the decisions, not the data. Write the recurring decisions down, map each to one observable field, freeze the competitor set, then build the change tiles before any trend charts.
Five steps, about half a day for a first working version.
- Write down the recurring decisions first. List the choices your team makes on a rhythm: what to publish next, whether to respond to a competitor move, where to put the ad budget, what to fix in the product story. Five to eight is normal.
- Map each decision to one observable field. For every decision, name the single public field that would change your mind. If no observable field exists, the decision belongs in a research task rather than on a dashboard, and pretending otherwise produces a tile nobody trusts.
- Freeze the competitor set before building anything. Pick three to six competitors and write the list down with a date. Every trend on the dashboard is computed against that set, so a set that changes mid-quarter invalidates the entire history behind it.Record why each competitor is in the set. That note is what makes a later removal defensible.
- Build the change band before the trend band. Changes since the last review are what get acted on, so they go at the top and get built first. Trend charts are for the monthly and quarterly conversation, and a dashboard with only trends produces admiration rather than decisions.
- Attach a name and a review slot. One person owns the screen and one recurring meeting reads it. A dashboard with no scheduled reader decays silently, because nobody is present at the moment a source quietly stops returning data.
4. How many competitors and how many metrics should it hold?
Three to six competitors and around six metrics each. The number that decides whether the dashboard survives is cells to read per review, and it rises as a product of the two.
| Tracked competitors | Metrics each | Cells to read per review | Survives contact with a real week? |
|---|---|---|---|
| 3 | 6 | 18 | Yes, about ten minutes |
| 5 | 6 | 30 | Yes, about twenty minutes |
| 8 | 10 | 80 | Rarely, the review starts getting skipped |
| 12 | 12 | 144 | No, unread by the third week |
Tile count and set size are separate decisions. Most tiles should aggregate across the whole set, showing one row per competitor inside a single tile, so growing the set from three to five adds rows rather than tiles. The screen stays at eight tiles and the reading time grows in a way you can predict.
Refresh rate is where cadence data changes the design. In the Oppira Benchmark for June 2026, tracked accounts averaged 1.98 posts per week on Instagram, 1.68 on Facebook and 1.04 on LinkedIn, which means a daily posting tile is mostly showing zeros.
1.04posts/week
Typical LinkedIn cadence for a tracked competitor account
Oppira Benchmark, unweighted mean across 56 tracked LinkedIn accounts, as of June 30, 2026.
A tile that is empty on most days trains people to stop looking at it, and it will be empty on the day it matters too. Weekly is the right granularity for cadence, with daily collection underneath so the weekly number is complete.
5. Should the dashboard show our own numbers next to competitors?
Yes, and computed exactly the way the competitor numbers are computed. Pulling your figures from your own analytics while pulling theirs from public data compares two different formulas and calls the difference performance.
Your analytics give you reach and impressions. Public competitor data gives you followers and visible interactions. If your row on the dashboard uses the richer denominator and their rows do not, every comparison on the screen is wrong in your favour, which is the most dangerous direction for it to be wrong in.
The fix is to compute a second version of your own numbers using only the fields you can see for them. Keep the richer internal version for content decisions and put the like-for-like version on the competitor dashboard. Two numbers for the same thing is fine as long as each is labelled with the question it answers.
Your row also belongs in the set for any share calculation. Share of the tracked set is only meaningful when you are inside the set, and it is one of the few figures on a competitor dashboard that a leadership audience reads correctly without explanation.
6. How do I know whether the dashboard is working?
Count the decisions it changed. After three reviews, every tile should have either informed a choice or been formally retired, and a screen where nothing has ever changed a decision is decoration.
Four checks, run once a quarter, in writing:
- For each tile, name a decision from the last three reviews that it informed. Tiles with nothing to name get removed.
- For each decision on your original list, name the tile that feeds it. Decisions with no tile are the real gap.
- Check the last successful collection date on every source. A source that broke quietly is the most common cause of a dashboard that looks calm.
- Ask which tile people look at first. That is usually the one worth expanding, and the answer is rarely the one that took longest to build.
7. Why do competitor dashboards go stale, and how do I stop it?
Three causes: the competitor set stops matching the market, a metric definition drifts without anyone restating history, and a data source fails silently. All three are preventable with a date and a note.
Set drift is the slowest and the most damaging. Competitors get acquired, pivot away from your category, or stop being the ones you lose deals to, and a dashboard tracking the market of eighteen months ago produces confident answers to a question nobody is asking. Review the set quarterly and record every change with a reason and a date.
Definition drift is quieter. Someone adds saves to the interaction count, or switches from per-post averages to period totals, and the trend line steps in a way that looks like a competitor result. Any change of definition means restating the whole series, which is cheap if the raw data is kept and impossible if only the aggregates are.
Silent source failure is the one that produces wrong decisions fastest, because an empty tile reads as a quiet competitor. Put the last successful collection timestamp on the dashboard itself, next to the data rather than in a settings page. Oppira collects each tracked channel daily and keeps the change history, so a gap in the record shows up as a gap rather than as calm.
Key Takeaways
One tile per recurring decision
Before adding a tile, name the question it answers and the person who asks it. Data availability is not a reason to display something.
Eight tiles covers most small teams
Changes, cadence, ad count, ad messaging, pricing and page changes, review movement, share of set, and your own numbers on the same fields.
Reading work is a product, not a sum
Five competitors at six metrics is thirty cells per review. Twelve at twelve is 144, which is why large dashboards go unread by week three.
Keep estimates off the screen
Modelled ad spend and traffic estimates lose their caveats within a week of appearing on a dashboard and then get quoted as observations.
Match refresh rate to how fast things move
Tracked accounts averaged 1.04 posts per week on LinkedIn in June 2026, so a daily cadence tile shows mostly zeros and trains people to ignore it.
Show the last collection date on the dashboard
A source that fails silently makes an empty tile look like a quiet competitor, which is the fastest route to a confidently wrong decision.
Frequently Asked Questions
Explore More
Related analyses, benchmarks, and industry insights
Related Guides
Glossary Terms
Turn reading into a reaction
Oppira watches your competitors, keeps your playbook current, and drafts the response. Start free and see your market clearly by tomorrow.