The three metrics that predict whether your content works

May 7, 2026

The three metrics that predict whether your content works

A typical content dashboard ships with two dozen metrics. Three of them predict whether a post earned its slot in the calendar. The rest produce anxiety after the fact and change nothing about next week. This article is for content directors and brand operators who want a dashboard that drives decisions rather than retrospectives, and the three metrics below are the only ones that survive the cut.

The standard failure pattern is a Monday morning review where the team scrolls through impressions, reach, total likes, comment count, and follower growth, and concludes that "engagement was up". None of those numbers tell anyone what to ship more of next week. They are downstream of the three metrics that matter, and watching them in isolation is post-hoc rationalisation dressed up as analysis.

The three metrics that predict whether content works

The three metrics are time to first 100 views, save rate, and downstream action. Each of them predicts something the dashboard cannot recover after the fact. The first predicts the algorithmic ceiling on the post; the second predicts whether the audience will return; the third predicts whether the post earned its slot against the cost of producing it. The rest of the dashboard is downstream noise.

Use the three together, not individually. A post can win on time-to-100 and lose on save rate, which means the algorithm rewarded a hook the audience did not actually value. A post can win on save rate and lose on downstream action, which means the audience valued it but it did not move the brand. The interesting decisions sit at the intersections, not the maxima.

Metric one: time to first 100 views

The metric is the time from publish to first 100 views, measured in minutes. Total view count and engagement rate are downstream of it; the time-to-100 number predicts a post ceiling within the first hour. Every major platform algorithm front-loads distribution to early performers. A post that breaks 100 views inside the first hour is on a fundamentally different trajectory than one that takes six hours to get there. Slow starts almost never catch up. Read the curve, decide whether to boost or move on.

The operational use of time-to-100 is binary in the first sixty minutes. Either the post crossed the threshold and is worth amplifying with a comment-pin, a story share, or a paid boost, or it did not and the budget belongs on the next post. Watching the curve at minute thirty changes nothing; deciding at minute sixty changes the next month of distribution.

Metric two: save rate

A like is the cheapest action a viewer can take, and a save is the most expensive one. A save means the viewer wants to come back to the post, which is the only durable signal that a post earned its place in the feed. On Instagram, save rate above 1.5 percent of impressions reads as healthy. Above 3 percent reads as excellent. Below 0.5 percent means the algorithm gave the post distribution but nobody wanted to remember it. Optimise for save rate, and the like count usually takes care of itself.

Save rate is the closest social-platform proxy for return-visit intent. A save reveals a viewer who plans to use the post later, which is the same intent that drives newsletter signups, pinned bookmarks, and direct-to-site traffic. Treat save rate as the brand-credibility metric on the content side, and you will start choosing topics that earn saves rather than topics that chase reach.

Metric three: downstream action

Did the post drive a measurable next step? A site visit, a newsletter signup, a booked call, a direct message. The exact conversion event matters less than measuring something concrete. UTM links plus a 30-day attribution window get you 90 percent of the way there. The interesting move is to plot conversion against engagement. The ranking inverts more often than you would expect. The teaching post with 50 saves and 2 booked calls quietly outperforms the viral hot-take with 5,000 likes and zero conversions. The hot-take feels like the win in the moment; the teaching post is what compounds.

Downstream action is the only one of the three metrics that is brand-specific. Time-to-100 and save rate generalise across platforms; downstream action depends on what the brand is selling, who it is selling to, and how the funnel is wired. Choose one downstream action per platform and stick with it for at least a quarter; switching the conversion event mid-quarter destroys the comparison.

The framework: three columns and how they replace the rest

Time-to-100. First column. Numerical value in minutes. Sortable. Color-coded green if under 60, amber if 60 to 240, red if over 240. The color is the call-to-action: green means amplify, amber means watch, red means move on.

Save rate. Second column. Percentage of impressions. Color-coded green if over 1.5 percent (over 3 percent is excellent and worth a flag), amber if 0.5 to 1.5, red if under 0.5. Save rate is the durable signal; over a quarter the green-column posts reveal the topics the audience values.

Downstream action. Third column. Count of the chosen conversion event in the 30-day window post-publish. The column does not need a color because the absolute number is the point: a quarter of green-on-conversion posts is the working content engine, and the brand should produce more of whatever they share.

A fourth column is optional and contains one line of post context (pillar, hook type, format). It is not a metric; it is a tag that makes the three-column dashboard sortable by hypothesis. Filtering by pillar reveals which pillars compound. Filtering by hook type reveals which hooks earn saves. The tag column is where the next-week decision is made.

Runbook: how to install the three-metric dashboard in week one

1. Strip the existing dashboard down to three columns. Impressions, reach, total likes, comment count, and follower growth come off. Resist the temptation to keep them as "secondary metrics"; secondary columns become primary inside a quarter. 2. Pull time-to-100 from the platform-native analytics. On Instagram and TikTok the curve is exposed natively; on LinkedIn and X you read the audience curve and mark the 100-view crossing manually for the first month, then automate. 3. Pull save rate from the native field. Every major platform that supports saves exposes save count and impression count; save rate is one division. If the platform does not expose saves, use the closest equivalent (replies on threads, reposts on X with a caveat). 4. Wire downstream action to UTM tracking and a 30-day attribution window. The window matters because content compounds; a 7-day window will under-credit the teaching posts that drive conversion in week three. 5. Run the three-column dashboard for four weeks before drawing conclusions. The first two weeks are noise and the last two are signal; quarterly the signal stabilises and the topic-level patterns become unambiguous. 6. Share the dashboard with the drafter and the reviewer. The decision-makers who write and approve content need to see the same numbers as the strategist; otherwise the dashboard becomes a separate artefact and the team writes against the old metrics. 7. Run the post-quarter retro on the three columns only. Topics that win on save rate and downstream action get more cadence next quarter. Topics that win on time-to-100 alone get re-examined; algorithmic distribution without retention is a sugar high.

When the three-metric dashboard is the wrong tool

The three metrics under-serve brands that are running pure brand-awareness campaigns with no conversion event. If the brand is genuinely not measuring downstream action because the goal is reach, downstream action will read as zero on every row, and the dashboard collapses to two columns. That is fine, but the brand should be honest that no slot in the calendar is being earned by ROI; it is being earned by perceived reach, which is the metric the rest of the dashboard already shows.

The metrics are also wrong for very high-volume cadences where time-to-100 is so consistent across posts that the column ceases to discriminate. Brands posting forty times a week often see time-to-100 cluster tightly because the algorithm has learned the account; in those cases save rate and downstream action carry the full weight, and time-to-100 becomes a health check rather than a decision metric.

They are wrong, finally, for new accounts in their first ninety days. The algorithm has not yet learned the account, save rates are unstable, and downstream action is too small a sample to read. Use the dashboard as a discipline-building exercise during the first quarter, but draw conclusions only after the account has reached steady-state distribution.

What success looks like with the three-metric dashboard

A team running this dashboard makes sharper next-week decisions inside one month, identifies its three most-compounding topics inside one quarter, and earns measurably more downstream action per post by month six. The save-rate column tightens as the team learns what the audience values; the time-to-100 column reveals which hook types get the algorithm to distribute; the downstream action column reveals which topics actually move the brand.

The same three-metric discipline applies on the retention side: revenue per inbox-placed send, save-rate-equivalent (which is reply rate for cold and click-rate for warm), and downstream conversion. The 90-day Retention Architecture on /email-and-sms is built around those three numbers, not the forty in the default Klaviyo dashboard. The qualitative signal is the same in both cases: the weekly review takes ten minutes instead of an hour, and the decisions are about what to do, not what to feel.

FAQ

What if my platform does not expose save rate natively? Use the closest equivalent. On X, reposts function as a save proxy; on LinkedIn, the share count carries similar intent. Reply rate is also a defensible substitute on text-heavy platforms. The principle is to find the most expensive action a viewer can take that signals return intent.

How do I handle posts with no clear downstream action? Pick a generic downstream action that applies to the whole calendar (a site visit, a newsletter signup) and use it as the default. Posts genuinely designed for awareness will read as zero on conversion; that is information, not a flaw, and the dashboard is honest about it.

Is time-to-100 still relevant on platforms that have moved away from chronological feeds? Yes. Algorithmic platforms still use early performance as a distribution signal; the chronological feed is what changed, not the front-loading. Time-to-100 measures the algorithm-perceived early performance, which is the input the platform actually uses to decide whether to broaden distribution.

Can I add a fourth metric without breaking the discipline? Be careful. The whole point of three columns is the cognitive forcing function. A fourth metric usually returns the dashboard to vanity. The exception is a tag column that holds context (pillar, hook type, format), which is not a metric but a sortable filter on the three.

How does this dashboard compose with the rest of the marketing system? The three columns are the content side of the same operator surface that runs retention and CRO. The save-rate-equivalent and downstream-action numbers compose across channels, which is why the AI-native marketing OS uses one dashboard rather than four.

- The retention three-metric mirror on Klaviyo: https://www.arthea.ai/email-and-sms - How an AI-native marketing OS unifies the dashboards: https://www.arthea.ai/article/ai-native-marketing-os - The five-step brief-to-ship process that produces the posts being measured: https://www.arthea.ai/article/5-step-brief-to-ship-process

If you want a 30-minute architecture review on the metrics layer for your brand, the calendar is here: https://www.arthea.ai/book.

Go back to Blog

Download the ressource