An SEO roadmap is a list of changes you’re going to make to your website, ranked by how much each one should help the business, with a month and a name next to every line. That’s all it is. No 40-slide deck required.
The part people get wrong isn’t the plan. Most roadmaps look fine in a spreadsheet, and then nothing on them goes live. Or a few things go live, nobody writes down which ones, and three months later nobody can say what worked.
A good roadmap does two jobs. It decides what to do next, and it keeps a record of what actually happened.
I put together a free Google Sheet that does both. It scores every task for impact, confidence and effort, warns you when you’ve planned more work than you can ship in a month, and has a changelog and a monthly review tab so the plan stays connected to the results.
You can make your own copy here. It comes with a filled-in example, so you can see how it works before you clear it out.
The short version
- An SEO roadmap is a ranked, dated list of changes to specific pages, with an owner for each one. Strategy decides where you’re going. The roadmap decides what changes on the site, and when.
- Start with the pages that make money, and save a baseline of where they stand before any work begins.
- Collect every possible change in one list, and write each one as a change to a specific URL, with the evidence behind it.
- Score every item for impact, confidence and effort. Then move anything that blocks other work to the top, which is usually tracking and indexing.
- Plan the next 90 days in detail and keep months four to twelve rough. They’ll change anyway.
- Only schedule what you or your team can ship in a month.
- Log every change the day it goes live, and review the numbers once a month.
What an SEO roadmap is (and what it isn’t)
People use roadmap, strategy, plan and audit as if they all mean the same thing, and that’s a big part of why so many roadmaps end up as long slide decks nobody opens twice. Or at all.
An SEO strategy is the big decision. Which customers you’re going after, which searches matter to them, and what kind of site you need to win those searches.
It changes once or twice a year, if that. An audit is a list of what’s wrong or missing on the site right now, and it’s one of the things a roadmap gets built from.
A content calendar is the schedule of what you’re publishing, which is one slice of the roadmap.
The roadmap sits between all of them. It takes the strategy, the audit and everything else you know about the site, decides what’s worth doing, and puts it in order with dates and owners.
You’ll also see it called an SEO plan, an SEO action plan or an SEO project plan. Same thing, different job titles.
| What it answers | How often it changes | |
|---|---|---|
| SEO strategy | Where are we going, and why? | Once or twice a year |
| SEO audit | What’s wrong or missing on the site right now? | Every time you run one |
| Content calendar | What are we publishing, and when? | Weekly or monthly |
| SEO roadmap | What changes on which page, in what order, and who does it? | Every month |
What’s in the template
The template is a Google Sheet with six tabs.
Open the copy link and Google will save a copy to your own Drive.
If you don’t use Google Sheets, download your copy as an Excel file with File > Download. Every tab has example rows for a made-up kitchen remodeling company, which you overwrite when you start.
| Tab | What it’s for |
|---|---|
| Start here | Short instructions and the scoring guide |
| Goals & baseline | Your business goal, the pages that make money, and where they stand today |
| Backlog | Every change you could make, scored for impact, confidence and effort |
| Timeline | What’s planned for each month, and whether you’ve planned more than you can ship |
| Changelog | Every change that went live, the date it went live, and when to check the result |
| Monthly review | The numbers for each month next to what shipped, and what you decided to do next |
Most SEO roadmap templates cover the first four in some form.
The last two are the reason I built my own. On client work I log everything I ship, and I keep a running list of what I recommended next to what actually got implemented, because without those two lists there’s no way to connect a change in rankings to anything you did.
A roadmap without a changelog is a wish list.
Step 1. Write down what the site is for
Before you list a single task, write one sentence about what the website is supposed to do for the business. More quote requests. More sales from one product line. More demo bookings.
It sounds obvious, but it’s what every later decision gets measured against, and if you skip it, every task on the list will look equally important. Which is how you end up spending a week on alt text.
Next, list the pages that make money. For a service business that’s usually the service pages, the page with the quote or booking form, and maybe a pricing page. For a store, it’s the categories and product pages. Next to each one, write the search it should show up for.
Not necessarily every keyword it could rank for, just the main one.
After that, save a baseline. Open the Performance report in Search Console, set it to the last three months, and write down impressions, clicks and position for each money page.
In GA4, note how many conversions came from organic search. That needs your form submission or purchase marked as a key event, and if it isn’t, that’s the first item on your roadmap. I go through which numbers matter and where to find them in my guide to measuring SEO.
Without a baseline, the monthly review later turns into guessing. With one, you can point at a page and say it went from position 14 to position 6 after the March rewrite.
The faster way, with Claude and an MCP
Copying numbers out of Search Console page by page gets old quickly. If you use a Search Console tool that has an MCP server, you can connect it to Claude and ask for the whole baseline in one go. ChatGPT works with the same servers if that’s what you use.
The two I’d look at are SEO Gets and SEO Stack. Both plug into Search Console and GA4, and both have an MCP connection for Claude and ChatGPT. I run Search Console through SEO Gets with Claude connected to it, so this is roughly what I’d ask.
Using my Search Console and GA4 data, give me clicks, impressions, average position, the top query and organic conversions for /kitchen-remodeling/, /kitchen-remodeling-cost/ and /projects/, for the last three months compared with the three months before.
Paste the answer into the Goals & baseline tab and you’re done. You can also ask Claude which of your pages get impressions for searches with buying intent, which is a decent way to check whether your list of money pages is missing anything.
The tools themselves help with the rest of the roadmap too. In SEO Gets you can put your money pages into a content group, so the baseline and every monthly check after it are one click instead of five filters.
SEO Stack keeps your Search Console data for years once it’s connected, while Search Console itself deletes anything older than 16 months. That matters if your business is seasonal and you want to compare this spring with the last two.
SEO Gets is the cheaper one and has a free plan. SEO Stack costs more and has no free plan, but you get the longer history and an AI assistant built in.
Either way, check the numbers Claude gives you against Search Console the first time. AI is great at fetching data and sometimes a little too creative at summarizing it.
Step 2. Collect everything that could be done
With the goal and the money pages written down, everything you might do goes into the Backlog tab, one change per row.
The rule that matters most is to write each row as a change to a specific page. “Improve content” isn’t a task. “Add price ranges and a cost FAQ to /kitchen-remodeling-cost/” is.
The same goes for the Why column. Put the evidence there, like the page’s current position and impressions or the error a crawl found, so that when you score the row later you’re scoring facts and not vibes.
The ideas usually come from the same handful of places, and roughly in this order of payoff.
- Pages that are almost there. Filter the Search Console Performance report to pages sitting between positions 5 and 20 with a decent number of impressions. These are the fastest wins on almost any site. In Backlinko’s click-through study of about 4 million Google results, only 0.63% of searchers clicked anything on the second page, so a page at position 12 gets close to nothing and the same page at position 6 doesn’t. SEO Gets has a Striking Distance report that pulls this list for you, and if Claude is connected, you can ask it for the same thing (more on that in where AI helps in SEO). A Search Console export sorted by position does the job too.
- Anything blocking Google. Pages that should be indexed and aren’t, redirect chains, a stray noindex tag, broken canonicals. The Page indexing report in Search Console shows most of it, and my technical SEO checklist for WordPress covers the rest.
- Money pages that don’t match the search. Titles that don’t say what people type, thin service pages, missing prices, pages with almost no internal links pointing at them. My on-page SEO checklist is the list I work through, and internal linking has its own guide.
- Pages that don’t exist yet. Services you sell with no page, categories that don’t match how people search, questions customers ask on every sales call. Keyword research helps you find them.
- Pages fighting or fading. Two pages competing for the same search, or pages that used to bring in clicks and have been sliding for months. SEO Gets has Cannibalization and Content Decay reports for exactly these.
- Links and mentions. Suppliers, partners, the industry directories people use, local press. The methods worth your time are in how to get backlinks.
- Local work, if you serve an area. Google Business Profile, reviews and location pages, all covered in local SEO for small businesses.
What you shouldn’t do is paste an SEO tool’s site audit straight into the backlog.
A crawler will flag hundreds of warnings, most of them harmless, and if they all go in as separate rows they’ll bury the ten or fifteen things that matter. Read the audit, keep the findings that affect pages you care about, and add those.
Step 3. Score every item
A long backlog with no order is just a longer version of the problem, so every row gets three scores from 1 to 3. This is adapted from the ICE model that Sean Ellis came up with for ranking growth experiments, which multiplies impact, confidence and ease on a 1 to 10 scale.
I divide by effort instead and keep the scale at 1 to 3, because with ten options everyone spends 20 minutes arguing whether something is a 6 or a 7, and then it’s lunch.
| Score | Impact | Confidence | Effort |
|---|---|---|---|
| 3 | Affects a page that makes money, or something that blocks Google from many pages | You have evidence, like Search Console data or a confirmed error | A week or more, or it needs a developer |
| 2 | Supports a money page, like internal links or a guide that feeds it | A reasonable bet based on what ranks today | A few days |
| 1 | Nice to have, far from a sale | A hunch | An afternoon or less |
The priority score is impact times confidence divided by effort, so it runs from 0.3 to 9, and the template calculates it for you.
Try sorting the backlog by score and the top of the list is usually obvious. A quick fix on a money page with solid evidence behind it lands at 9. A big rebuild of a page that might help lands around 1 or 2.
Confidence is the score people skip, but it’s the exact one that actually makes roadmaps work.
A task that scores 3 for impact and 1 for confidence is a guess, and it should lose to a smaller task you know will work.
Most of the busywork that clogs SEO plans scores low on exactly this, especially publishing a blog post a week on topics nobody was searching for.
When Ahrefs analyzed about 14 billion pages in 2023, 96.55% of them got no traffic from Google at all, and the main reasons were no search demand, no backlinks, and content that didn’t match what searchers wanted.
That’s a whole internet of blog posts talking to nobody. Those are confidence problems, and you can usually see them before you write a word.
The score isn’t the final word, though.
Some tasks have to happen before others no matter how they score. If conversions aren’t tracked, nothing else can be measured, so tracking goes first.
If Google can’t index a section of the site, rewriting the content in it is pointless until that’s fixed.
That’s what the Depends on column is for. Put the ID of the blocking task there, and schedule the blocker first.
Step 4. Put it on a calendar you can keep
Scores tell you what matters most. They don’t tell you who’s doing it or when, and that’s where most roadmaps quietly fall apart.
So start with capacity. Who’s doing the work, and how much can they get done in a month?
If you run your own WordPress site in between everything else (so, at 11pm), that might be two or three small changes.
If you have a developer for a few hours a month, the technical rows go in their schedule, not yours.
The Timeline tab adds up the effort points planned for each month and compares them with the capacity you set, and a red cell means you’ve planned more than you can ship.
This matters more than it sounds.
In the 2023 State of Technical SEO report from Aira and Women in Tech SEO, 88% of respondents said they’d waited anywhere from a month to more than two years for technical SEO changes to be made, and 11% had waited over 24 months.
That’s the most depressing number in the whole report, and it has nothing to do with strategy.
Those plans were probably fine. They just never got a slot on anyone’s calendar.
That’s why you should plan the next 90 days in detail, with a month and an owner for every row.
Months four to twelve can stay rough, closer to themes than tasks (“build out the bathroom services section” instead of five separate page tasks), because by month three you’ll know things you don’t know now.
Some changes will have worked faster than you expected, and some won’t have moved at all.
The order usually follows the same logic. Fix what blocks Google, improve the pages that are close, build what’s missing, then earn links to the pages that matter.
I laid out what that looks like month by month in the how long SEO takes guide I wrote last month.
On a WordPress site, a lot of the roadmap is work you can do yourself. Titles, headings, content, internal links, redirects through your SEO plugin and most image fixes happen in the dashboard. This is one of the reasons I love WordPress.
Template changes, speed problems and anything that touches the theme’s code usually need a developer.
Mark those rows and get them into the developer’s schedule early, because they’re the ones that end up waiting.
Step 5. Log what ships and review it every month
The plan is the easy half. Every time something on the roadmap goes live, add a row to the Changelog with the date, the page, what changed and what you expect to happen.
The template fills in a check-again date six weeks later, because Google needs time to recrawl the page and settle on where it ranks. Judging a change after four days tells you nothing, except that you’re impatient. (Guilty.)
Add the same note in Search Console.
Google added custom annotations to the Performance report in November 2025, so you can mark the day a change went live right on the chart. Each note can be up to 120 characters, you get 200 per property, and everyone with access to the property can see them.
They do the job, but they’re basic. A note on a chart, and that’s it.
SEO Gets and SEO Stack both go further, and this is where they earn their place. SEO Gets has a Chrome extension for adding annotations while you’re on the page, it tracks content changes on your pages automatically, and it has an SEO testing feature for before and after comparisons. It’s where I annotate everything I ship.
SEO Stack lets you log the work, tie tasks to annotations and see the change on the chart when it lands, and it shows Google’s algorithm updates on the same chart. Its AI assistant will even tell you which annotations helped and which hurt.
And because both connect to Claude over MCP, you can skip the chart-squinting and ask something like “What happened to clicks and position for each page I annotated in the last two months, compared with the four weeks before the change?” The Changelog tab still matters either way. It’s the one place where the plan, the work and the results sit next to each other.
Once a month, open the Monthly review tab and give it fifteen minutes. Fill in clicks, impressions and conversions for the money pages, and the template counts how many changes shipped that month from the Changelog. Then write down what moved, what stalled, and what you’re changing next month as a result.
Every quarter, re-score the backlog.
New rows will have come in, some scores will look wrong now that you have data, and a few items will have stopped mattering. Delete those.
An example roadmap for the first 90 days
The example in the template puts all five steps on one small site, a made-up kitchen remodeling company in a mid-size US city.
It has a WordPress site of about 25 pages, an owner who can edit content, and a freelance developer for a few hours a month.
Capacity is set to 4 effort points a month.
| # | Change | Score | Month |
|---|---|---|---|
| 1 | GA4 key event for the quote form, Search Console linked to GA4 | 9.0 | Month 1 |
| 2 | New title and H1 on /kitchen-remodeling/ (position 9.2, 2,400 impressions) | 9.0 | Month 1 |
| 3 | Fix redirect chains on the old /kitchens/ URLs | 6.0 | Month 1 |
| 4 | Add real price ranges to /kitchen-remodeling-cost/ (position 14) | 3.0 | Month 2 |
| 5 | Link 12 kitchen design posts to the main service page | 4.0 | Month 2 |
| 6 | Build a bathroom remodeling service page | 3.0 | Month 3 |
| 7 | Add 8 recent projects to /projects/ | 2.0 | Month 3 |
| 8 | Get listed on suppliers’ installer pages | 2.0 | Month 4 |
| 9 | Compress gallery images on /projects/ | 1.0 | Later |
| 10 | Write 10 blog posts on kitchen design trends | 0.3 | Dropped |
Tracking goes first even though it doesn’t move a single ranking, because every other row needs it to be measured.
The title rewrite scores a 9 because the page already sits at position 9 with real impressions and the fix takes an hour.
The cost page scores lower than the internal links, but both fit into month two, so the order inside the month barely matters.
The image compression scores a 1 because the page isn’t a ranking target yet, and it depends on the new projects going up first.
Month three is exactly full. If the owner tried to squeeze the supplier outreach in there too, the Timeline tab would turn red, and that’s a conversation better had before the month starts.
The blog posts got dropped.
Ten articles on design trends sounds productive, and it would look great in a monthly report, but nothing suggests the people browsing kitchen trends at midnight are about to hire a remodeler in this city.
At 0.3, it would have eaten a month of writing for nothing.
Why most SEO roadmaps die in a spreadsheet
Most roadmaps don’t fail because the plan was bad. They fail in boring, predictable ways.
- Everything is high priority. If every row says high, the roadmap isn’t deciding anything. That’s what the scores are for.
- Nobody owns anything. A row without a name next to it doesn’t ship. “The team” isn’t an owner. “The team” is where tasks go to retire.
- It’s an audit export. Four hundred rows of crawler warnings with no evidence and no page attached. Nobody can work through that, so nobody does.
- The tasks can’t be finished. “Improve E-E-A-T” and “build topical authority” never get checked off. Break them into changes to specific pages that one person can finish in a week.
- It’s planned in detail for twelve months. That far out, a detailed plan is fiction, and it makes the roadmap too painful to update when month three goes differently.
- Nobody opens it again. A roadmap that only gets looked at for the quarterly report is decoration.
If an agency is building it for you
Everything above works the same when someone else does the work, and it gives you a very easy way to check on them.
If you’re paying someone for SEO, ask to see the roadmap, and ask to see it in this shape.
Specific pages, specific changes, a month for each, and the person doing it. If what you get is a strategy deck with pillars and phases and no URLs, you’ve paid for a very nice PDF. You can’t hold anyone to it, and you can’t tell whether the work happened.
Ask for the changelog too. Google’s own advice on hiring an SEO suggests asking “Will you share all changes made to my site?”, and a roadmap plus a changelog is that answer in writing.
If the changelog stays empty for two months in a row, the roadmap isn’t your problem.
If you want more of this kind of thing, I write a newsletter called Built to Rank. It’s occasional emails on what’s actually working in SEO, written from my real day-to-day work, and it only goes out when I have something worth sending.
A couple more questions
How far ahead should an SEO roadmap go?
Six to twelve months, with only the first three planned in detail. Anything further out works better as themes, because what you learn in the first few months will change the rest.
How often should you update it?
Add to the changelog every time something ships, review the numbers monthly, and re-score the whole backlog once a quarter. Also revisit it after anything big, like a redesign, a migration, a new service, or a Google core update that moved your rankings.
Can AI build the roadmap for me?
It can do a lot of the legwork. With Claude connected to Search Console through SEO Gets or SEO Stack, it can pull the baseline, find pages sitting at positions 5 to 20, spot pages that are losing clicks, and suggest scores for each row. What it can’t do is know your margins, which services you want more of, or that your developer is on vacation in August. Let it fill the backlog and do the first pass on scores, then make the calls yourself.
What is the 80/20 rule in SEO?
It’s the idea that a small share of your pages and tasks produces most of the results. Open the Pages tab in the Search Console Performance report for almost any site and you’ll find a handful of URLs bringing in most of the clicks. For a roadmap, it means spending most of your time on the money pages and the few changes that score highest, instead of spreading effort evenly across everything an audit found.
Do local businesses need a different roadmap?
Same structure, different rows. A local business adds Google Business Profile work, reviews, and location or service-area pages to the backlog, and those often score high because they affect the searches that bring in calls.
Should AI search be on the roadmap?
Mostly it’s already on there. The pages that rank well, answer questions clearly and get mentioned by other sites tend to be the ones AI tools cite as well.
What’s worth adding is measurement. Check how visits from Claude, ChatGPT, Gemini and other AI assistants show up in your GA4, and add them as a line in the monthly review so you can watch the trend.
Tools that track how often AI answers mention your business do exist, but the data is still rough, so I wouldn’t plan the roadmap around them yet.