SEO Forecasting at Scale
what 15 years of estimating traffic has taught me
All the SEOs I know are working on forecasting and business planning right now, so I wanted to share some quick tips I’ve learned over the years.
A quick caveat here that I’m talking about forecasting at scale: projects that involve growing or creating a whole template type, building out content in a new topic area, etc. These forecasts are typically used for making major decisions like which projects to prioritize for the year or whether an effort merits additional headcount or agency support.
While I am a big advocate for individual page optimizations (tweaking a page title, improving a blog post to better match search intent), forecasting and making charts about exactly how much we could improve if we moved up a rank or two on individual keywords isn’t something I spend my time on. My workflow for that is just (1) notice a quick opportunity, (2) action on it, and (3) eventually report on the growth in a weekly or monthly recap. (In those cases, I’d rather do the work than waste time talking about how good it might be if I did.)
At this point in my career, I’ve made hundreds of business cases and forecasts. And when projects move forward, I take the time to check actuals vs our predictions. There are some common thresholds (like highest feasible market share) that show up over and over again, and I find they are often lower than what junior SEOs might suggest when they first begin creating business cases and approach forecasting. (If you are using CTR curves to forecast, this newsletter might be especially valuable for you.)
New Content and New Markets
When it comes to entering new markets or building out new pages (projects like tackling a new topic area on a blog, creating facets or subcategories for an e-commerce style business), the bulk of my forecasting work involves building out a good, representative body of keywords. Often, the final list includes many thousands of keywords.
Typically I approach building that keyword list as follows:
Use site data/text to create target keywords. For an e-commerce site, this might be “Color (which I scrape from the sidebar) + Category Name” (which I scrape from the H1). For travel, I might pull city names and mash them up with common search terms (“Minneapolis day trips”). If I was Netflix, I might scrape carousel names like “RomComs with a Strong Female Lead” to decide whether these groupings deserved their own indexable landing pages.
Use a tool like Semrush to source keyword ideas from competitors. When I take this approach, I get as specific as I possibly can:
I only include keywords where the competitor ranks in the top 10 to improve relevancy.
I limit to specific subfolders or scrape their site for categories and do a URL mashup to limit by a specific category or tag.
Crawl someone else’s site for ideas. If my site doesn’t have what I need, I’ll crawl and scrape keywords from a competitor site. For content projects, I sometimes crawl forums or other blogs for ideas.
Once I have my keyword list, I use a rank checker to pull search volume. If my list is big enough and I don’t have access to an enterprise tool, I’ll rank check in batches, swapping out my list of keywords each day.
That work will get me to a total monthly search volume for our target market. If I’m using Semrush competitor data, I leave the search volume alone. If I am using a keyword list that I have created from scratch using site data, where each page that will be created is represented by 1 keyword (like “blue sofas”), I double my search volume estimate to account for long tail traffic.
From there, the question is how much traffic can we get? The answer is, best case scenario, you’ll get 2-5% of search volume as traffic. In a few years, if you build a best-in-class brand and dominate the market, you can get up to 8%. (For context, NerdWallet gets 8% market share in credit cards, Netflix’s traffic is 9% of its total search volume.) There are, of course, occasional exceptions to this rule where a very dominant brand absolutely owns a market and can get to 18%, but I really commonly see an 8% threshold for successful, strong-brands-who-are-good-at-SEO sites.
This is often where I see junior SEOs fall back to CTR curves and make outlandish estimates by using CTRs and assuming an average rank of 1-3. The flaw here is that nobody, not even the best brands in the world, rank in the top 3 for every one of their keywords. They rank well for some keywords, worse for others, and not at all for the rest. CTR is a nice idealized solution, but in the real world, 2-5% is common.
The answer of whether you’ll get 2% or 5% depends on how much brand authority you have overall, how much topical authority you have in this vertical, and how well you execute the nuts and bolts of the project.
I’ll often pitch a low/medium/high estimate, including what I think is likely. For brands that are well established and authoritative, that 8% is a North Star guidance that we aim to hit in a few years.
Most Site Enhancements Get You 5-15% Growth
This is simply a truth that 15 years of consulting has taught me. Whether you’re adding unique text to product pages, improving page design, restructuring your architecture, or migrating a subdomain to a subfolder - the result is going to be somewhere between 5-15% growth of unbranded traffic, applied to the pages where the change is being made. (Of course, the devil is in the details.)
One exception to this is sites that are in poor health, with severe structural problems that are majorly impeding performance: things like content that can’t be rendered, millions of duplicate pages, JavaScript issues, etc. Of course, in those cases, where a site is severely under-performing, solving a major issue can unlock massive growth and easily improve traffic by 2X or 3X.
The other exception to this rule is start-ups in high-growth stages. In those cases, I’ll use market share as a target (what if we get to 5% or 8% of the total market search volume) instead of basing growth on current performance.
But, setting those exceptions aside for a minute, if someone came to me from a well established site and said “I think we could customize our vendor descriptions and double traffic,” I would be extremely skeptical. Can you stack projects to get to a point of doubling traffic? Sure. But rarely are there single optimization projects or site enhancements that can get you there on their own.
Here are some common pitfalls I see people run into when forecasting on-site projects:
Your growth estimates should (usually) only be applied to unbranded traffic. You likely already rank #1 for branded searches, and there’s no room to grow.
Your growth estimates should only be applied to pages that will be affected by the change. For example, if you add text to product pages, only product pages will lift. There are occasional projects that trigger a site wide lift, but I find those are often weighted so that some sections get a larger lift and the rest of the site gets a much smaller impact.
You need to use RPV (Revenue per Visit) for the relevant template type. I talked a lot about traffic estimates in this post, which is how I always begin my forecasts, but, for delivery, I translate everything into revenue. The metric I like to use for revenue opportunities is RPV. For 99% of sites, the blog/content RPV is much lower than a product or category page RPV, so it’s important to apply the relevant dollar amount to the project. It’s not uncommon for a lower traffic product page enhancement to beat out a high traffic content initiative, because it nets more $$$ for the business.
I’m Doing an AMA Tomorrow
Tomorrow morning, (Thursday, 2/14/2025), I’ll be doing an AMA in the SEO Community Slack. I quit Twitter (when it was still Twitter) years ago, and I have found this Slack to be a much more peaceful means of staying up on SEO news and trends and building community. My best pitch for it is that it’s a place where there are a lot fewer SEO hustlers trying to build their personal brands and a lot more people who actually work on websites sharing ideas and asking for advice.
As always, thanks for reading and sharing. I’d love to hear if you also find these “3-8% of the total market” and “5-15% growth of existing content” rules to be true for you.