SEO Forecasting: Honest Projections for Small Sites
SEO forecasting for low-authority sites requires different math. Learn a CTR-curve x volume x probability method that gives honest traffic projections.
TL;DR -- Most SEO forecasting models assume you will rank. On a low-authority site, the probability of ranking is the dominant variable, and most forecasts ignore it entirely. Use a CTR-curve times volume times probability-of-ranking method, be honest about the error bars, and you will get projections you can actually plan around.
You just built a content calendar. Twelve posts, clustered, sequenced by winnability. Your co-founder asks a reasonable question: how much traffic will this generate? You open a spreadsheet, multiply keyword volumes by some click-through rate, and arrive at 4,200 monthly sessions. The number feels real because it came from a formula. Three months later, you have 310 sessions. The forecast was not slightly wrong. It was off by an order of magnitude.
I have made this exact mistake. The formula was correct in structure but missing a variable that dominates everything else at low DR: the probability that your page actually reaches a ranking position where clicks happen. Without that variable, SEO forecasting is fiction dressed up as math.
SEO forecasting starts with the variable nobody includes
The standard SEO forecasting method works like this: take each target keyword's monthly search volume, multiply by the expected click-through rate for your anticipated ranking position, sum the results, and project monthly organic traffic. Agencies use this model. Enterprise teams use it. It is embedded in most SEO forecasting tools and templates.
The problem is the phrase "anticipated ranking position." On a DR-60 site targeting KD-15 keywords, anticipating a top-three position is reasonable. The site has the authority, the backlinks, and the indexing history to reach page one for moderate-difficulty queries. On a DR-8 site, that assumption is the entire question. You are not forecasting traffic. You are guessing whether you will rank at all.
An honest SEO forecasting model for a small site needs three inputs, not two:
| Input | What it represents | Where it comes from |
|---|---|---|
| Search volume | Total monthly searches for the keyword | Keyword tools (DataForSEO, Ahrefs, etc.) |
| CTR by position | Click share at each ranking position | CTR curve studies |
| P(rank) | Probability your page reaches that position | Winnability analysis against the SERP |
The formula becomes:
Projected monthly clicks = Volume x CTR(position) x P(rank at that position)
That third variable -- probability of ranking -- is what separates a useful forecast from a fantasy. It is also the variable that most SEO forecasting software quietly sets to 1.0, which assumes you will rank for everything you target.
The CTR curve: what the data actually says
Click-through rates by position are the best-documented piece of the forecasting equation. Multiple large-scale studies have measured them. According to a 2024 study by Backlinko analyzing over 4 million search results, the position-one organic result captures roughly 27.6 percent of clicks. Position two gets about 15.8 percent. Position three drops to around 11 percent. By position ten, you are at roughly 2.4 percent.
For a small-site forecast, the useful simplification looks like this:
| Position | Approximate CTR |
|---|---|
| 1 | 27% |
| 2 | 16% |
| 3 | 11% |
| 4-5 | 6-8% |
| 6-10 | 2-5% |
| 11-20 | Under 1% |
| 21 and beyond | Effectively zero |
These numbers shift based on SERP features -- a featured snippet absorbs clicks from position one, "People Also Ask" boxes compress the organic results, and AI Overviews are changing click distribution in ways Google itself is still measuring. But as baseline estimates for forecasting, they are serviceable. The important takeaway: if your page does not reach the top ten, the CTR is effectively zero. Position 15 and position 150 produce roughly the same traffic: none.
This is why optimizing for featured snippets and ranking in People Also Ask boxes matters disproportionately for small sites. Those SERP features let a lower-authority page capture clicks it would not get from a standard organic listing alone.
How to estimate probability of ranking
This is the hard part. No tool gives you a clean probability number for whether your specific domain will rank in the top ten for a specific keyword. But you can build a reasonable estimate from signals you already have.
The winnability framework I use in other posts -- and that Boomranq automates -- translates directly into a probability estimate. Five signals, each contributing to a rough probability:
Signal 1: DR gap. How far is your domain rating from the weakest site currently on page one? If you are within 10 DR points of the lowest-DR result, P(rank) goes up. If the gap is 30 or more, P(rank) drops toward zero. The keyword difficulty checker explanation walks through why this gap matters more than the raw KD score.
Signal 2: SERP composition. Forums, Reddit threads, and thin pages in the top ten mean Google has not found strong dedicated content yet. That raises your probability. Five deep, well-linked pages from high-authority sites means the SERP is locked. That lowers it.
Signal 3: Topical cluster support. A page published as part of a cluster with internal links from siblings and a pillar has a higher probability of ranking than an identical page published in isolation. Topical authority is the mechanism -- Google uses the surrounding content to gauge relevance.
Signal 4: Content quality gap. If the current top results are outdated, thin, or poorly matched to the query intent, your probability of outranking them rises. If they are comprehensive and fresh, it drops.
Signal 5: Keyword difficulty. The raw KD score, properly contextualized. At DR 8, a KD of 3 with favorable signals might yield a P(rank) of 0.6 to 0.8. A KD of 15 with unfavorable signals might be 0.05.
I convert these into a rough three-tier estimate. This is deliberately imprecise because false precision in forecasting SEO traffic is worse than honest imprecision:
| Winnability tier | Signal pattern | P(rank in top 10) |
|---|---|---|
| High | Low KD, forums in SERP, DR gap under 10, cluster support | 0.5 to 0.8 |
| Medium | Moderate KD, mixed SERP, DR gap 10-20, some cluster support | 0.15 to 0.4 |
| Low | Higher KD, strong SERP, DR gap over 20, isolated post | 0.02 to 0.1 |
These ranges are wide on purpose. If your SEO forecasting template says "we will get exactly 847 sessions from this keyword," the decimal precision is lying to you. A range of 200 to 600 is more honest and more useful for planning.
Building the forecast: a worked example
Here is how the three-variable model works in practice. Say you have a five-keyword cluster for a DR-10 SaaS site:
| Keyword | Volume | KD | Winnability tier | P(rank top 5) | CTR at position 3-5 | Projected clicks/mo |
|---|---|---|---|---|---|---|
| "sprint retro template for contractors" | 40 | 2 | High | 0.7 | 9% | 2-3 |
| "async retrospective tool" | 30 | 4 | High | 0.6 | 9% | 1-2 |
| "remote retro best practices" | 110 | 8 | Medium | 0.3 | 9% | 2-4 |
| "retrospective template agile" | 260 | 12 | Medium | 0.2 | 7% | 2-5 |
| "sprint retrospective guide" | 390 | 16 | Low | 0.08 | 7% | 1-3 |
Total projected range for the cluster: roughly 8 to 17 clicks per month at the three-month mark.
That number looks small. It is small. And it is honest. A model that ignores P(rank) and just multiplies volume by CTR would project about 75 clicks per month from this same cluster. The founder budgets time based on 75. The reality is closer to 12. The disillusionment that follows is not a failure of content marketing -- it is a failure of forecasting.
The honest forecast also reveals something the inflated one hides: most of your near-term traffic will come from the easiest keywords. This reinforces the winnability-first sequencing principle. Publish the high-probability keywords first, let them rank and compound, then tackle the medium-tier keywords with the authority you have built.
Why most SEO forecasting tools overproject
Most SEO forecasting tools are built for agencies pitching enterprise clients. They apply a CTR curve to a keyword list and produce a traffic projection that justifies a retainer. The probability-of-ranking variable is either set to one or hidden behind a vague "expected position" input the user fills in optimistically.
This is not malice. It is a mismatch of context. An agency working with a DR-65 site can reasonably assume top-ten placement for sub-30 KD keywords. But when a DR-8 founder downloads the same SEO forecasting template or plugs keywords into the same forecasting SEO software, the implicit assumption -- that ranking is likely -- no longer holds. The forecast overpromises by five to ten times. The founder concludes SEO does not work. SEO works fine. The forecast did not.
On top of the P(rank) gap, every input carries error bars. According to Ahrefs' own analysis of search volume accuracy, volume estimates can deviate substantially from actual searches, especially for low-volume keywords. CTR curves shift as SERP features and AI Overviews absorb clicks. Ranking timelines vary: the Ahrefs study on time to rank found only 5.7 percent of new pages reached the top 10 within a year. And Google changes the rules -- core updates can move your page from position 8 to 18 overnight.
Given all this, I treat any small-site SEO forecast as a range. If the model says 15 clicks per month, the honest version is 5 to 30, assuming the cluster is published in winnability order and refreshed based on reporting data that drives decisions.
Connecting forecasts to ROI
A traffic forecast alone does not answer the founder's real question: is this worth my time? Connect the forecast to a simple conversion funnel -- projected clicks times visitor-to-trial rate times trial-to-paid rate times average revenue per customer -- and you get a revenue range per cluster.
If your cluster projects 8 to 17 clicks per month, and your visitor-to-trial rate is 3 percent, and your trial-to-paid rate is 15 percent, you are looking at roughly 0.04 to 0.08 new customers per month. Not impressive in month one. But content compounds. By month six, if three posts have reached page one, the cluster might generate 40 to 80 clicks. That is where the SEO forecast connects to content marketing ROI measurement -- the forecast gives the expected range, the ROI framework tells you whether reality is tracking within it.
Turning the forecast into a template
The worked example above is the template. One row per keyword, one sheet per cluster: volume, KD, winnability tier, P(rank), CTR, and projected clicks as a low-high range. Sum the cluster. State it as a range. That is your forecast.
The template is simple because the hard part is not the arithmetic. It is the upstream work: identifying winnable keywords, clustering them, and estimating P(rank) honestly -- then having the discipline to write down numbers that feel disappointingly small.
The forecast is the plan
Here is the contrarian take on SEO forecasting that I keep coming back to: the winnability score in your content calendar is already a forecast. When you score a keyword as "high winnability" and slot it into week one of a 30-day calendar, you are making a probabilistic claim -- this keyword has a high chance of reaching page one given our domain's current profile. When you sequence keywords from highest to lowest winnability, you are forecasting that early wins will compound into later ones.
Most founders think of forecasting SEO traffic as something separate from planning -- a projection you make after the calendar is built. But a winnability-first calendar is a forecast expressed as a publishing schedule. Each slot says "we believe this keyword is winnable at this point in time, given everything we will have published before it."
That is the model Boomranq builds when you give it a product description. It scores winnability, estimates the probability of ranking per keyword, clusters and sequences the calendar, and the output is simultaneously a plan and a forecast. The projection is not layered on top of the calendar. It is the calendar.
Whether you build that model in a spreadsheet or let a tool do it, the principle holds: forecast honestly, plan around the honest numbers, and measure against reality. The founders who succeed with SEO are not the ones with the best projections. They are the ones whose projections were honest enough to plan around -- and who kept measuring until the compounding math proved them right.