SaaS keyword map template

Most keyword research dies inside a 400 row export nobody opens twice. The map is the artifact that turns that export into assignments: one keyword, one URL, one owner, one due date. Everything below is the sheet structure, the two formulas that make it police itself, and twenty filled rows for a fictional $15K ACV revenue intelligence product so you can see what a finished row actually looks like.

The twelve columns, and who fills each one

Twelve columns is the point where the sheet stays readable and still carries enough to plan against. Fewer and you lose the value model. More and people stop updating it.

Columns 8, 9 and 12 are formulas. Nobody types into them. That matters more than it sounds, because the fastest way to kill a shared sheet is to let three people hand enter a priority number based on vibes.

Why one keyword maps to exactly one URL

This is the rule the whole template exists to enforce. One keyword owns one URL. One URL owns one primary keyword plus the variants that share its intent.

A comparison page can legitimately hold the plural, the year qualified phrasing and the vs reversal, because all three readers want the same page. It cannot also hold the definitional query for the category, because that reader wants a definition and will bounce off a product table. Assigning both to one URL is how you end up with a page that ranks eleventh for two things instead of third for one.

The mechanical payoff is that keyword cannibalization becomes a spreadsheet error rather than a diagnosis you make nine months later from a Search Console query report. When two rows claim the same target URL, the flag fires that day.

The modelled pipeline value formula

Column 8 is the one that changes publishing order. It replaces volume as the sort key.

Monthly sessions = Volume × CTR at target position
Monthly signups = Monthly sessions × Visitor-to-signup rate
Monthly opps = Monthly signups × Signup-to-opportunity rate
Annual pipeline = Monthly opps × 12 × ACV

For the worked example the inputs are: target position 3 with a 10% CTR on commercial queries and 6% on informational ones after AI Overview compression, ACV $15,000, and conversion rates that differ by intent tier. Those tier rates are the part everyone gets wrong by using a single site wide number.

A T1 keyword at 320 searches a month outvalues a T4 keyword at 9,000. Run the arithmetic and the gap is roughly 14x. That single fact is why the sort column matters, and it is the same logic behind building bottom of funnel SEO for SaaS pages before anything educational.

The priority score, and why it is not just value

Value alone pushes every engineering dependent page to the top, and those take three months. Divide by an effort band to get something a team can actually sequence.

Priority score = (Annual pipeline value ÷ 1000) ÷ Effort points

Effort points: 1 = writer only, no new assets
 2 = needs SME interview, screenshots or original data
 3 = needs engineering, a data feed or a new template

Sort descending on priority, then hand the top fifteen rows to the editor. Anything scoring under 5 goes back to the backlog tab. If you want the traffic side modelled properly before you commit headcount, run the set through the organic traffic forecast calculator and the SaaS SEO ROI calculator rather than trusting the sheet’s own straight line.

Twenty filled rows for a $15K ACV product

The company here sells revenue intelligence to RevOps leaders at Series B to Series D software companies. It competes with Gong on conversation data and with Outreach and Salesloft on workflow. Volumes are plausible planning figures, not lookups.

Total modelled annual pipeline across the twenty rows is roughly $1.14M. Look at what the sort produces. Apollo alternatives at 3,100 searches beats the 9,900 search definitional query by a factor of ten, and the three integration pages carry real value but drop down the queue because they need engineering. That sequencing decision is the map’s whole job.

The cannibalisation flag, in two formulas

Column 12 holds one formula that catches both failure directions. In Google Sheets, with keywords in column A and target URLs in column D:

=IF(AND(D2<>"", COUNTIF($D$2:$D, D2)>1), "DUPE URL",
 IF(COUNTIF($A$2:$A, A2)>1, "DUPE KEYWORD",
 IF(D2="", "NO TARGET", "")))

Conditional formatting on that column, red for DUPE URL, amber for NO TARGET. Then add a second check that runs against reality rather than the plan: pull the Search Console query export monthly, and flag any query where two URLs both received clicks in the same month. The sheet catches planning collisions. The export catches the ones that already shipped.

Status, owners and the weekly fifteen minutes

Column 10 takes a person’s name. Not Content Team, not Agency. A named human, because unowned rows sit for a quarter and then get rediscovered in a planning session as though they were new.

Status moves through six values: Backlog, Briefed, Drafting, Review, Published, Refresh due. The last one is the important one and almost nobody includes it. Set a formula that flips any published row to Refresh due when its position drops more than five places or twelve months pass, whichever comes first. That single column is what stops a map from becoming a record of things you did in 2026.

Run the review weekly, fifteen minutes, three questions. What moved from Drafting to Review. What flags fired. What changed position enough to reorder the queue. The cadence side of this is covered properly in lesson 3 on cadence, distribution and targets.

What this template does not solve

It will not tell you whether the cluster is worth entering. A map built on a category where your product does not fit will produce beautifully sequenced pages that convert at 0.1%. Work out coverage first, which is what lesson 2 on building the coverage map is for, then map keywords inside the clusters that survive.

It also handles programmatic page sets badly. Four thousand integration URLs do not belong as four thousand rows. Represent the whole set as one row with a pattern in the target URL column, then manage the specifics through the programmatic page brief template where indexation thresholds and kill criteria live.

The honest cost: a first build takes a focused day for 200 keywords, and the maintenance runs about an hour a week. Teams that treat it as a one time exercise get three useful months out of it before the position column goes stale and people quietly revert to arguing about topics in Slack.

Build it this week

Copy the twelve columns into a new sheet. Paste your existing keyword export into a Backlog tab, not the map. Promote only the rows where you can name the target URL today, add the three formulas, then sort by priority and brief the top ten. If you want the whole thing sequenced against a start date, the 90 day SaaS SEO plan slots this map into week one, and the keyword map lesson walks through the build with a screen share. The broader programme context sits in the SaaS SEO hub.