Brand Search Incrementality
Bidding on branded keywords: how to know if yours are actually worth it
Bidding on branded keywords only pays when the clicks are incremental. If you already rank first for your own name, some of that spend is buying traffic you had for free. One test tells you how much — and it takes about three weeks.
Your Brand — Official Site
yourbrand.com
You paid for this click
Your Brand — Official Site
yourbrand.com
You already owned this one
Search your own brand name and this is what most brands see: their ad sitting directly above their own top organic listing. The question isn't whether that looks good. It's how many of those top clicks you'd have gotten anyway.
By Billy McCall, Paid Search Lead at Parallel Path·Updated September 2026
The right question
Should we bid on our own brand?
What's our organic replacement rate?
Bidding on branded keywords is worth it only when the clicks are incremental. The organic replacement rate measures that: organic clicks gained divided by paid clicks lost when you pause brand ads in test markets. Below 40%, keep investing. Above 80%, stop.
Jump to
-
Why the pause test isn't a test
Turning brand off for a month and comparing to last month proves nothing.
-
The replacement rate and its thresholds
Three bands, three decisions: pause, reduce, or maximize.
-
How to run a brand incrementality test
Geo split, matched markets, and how long to hold the window.
-
When not to run it
Four conditions that will give you a number you can't trust.
Fast answers
The four questions people actually search
Should you bid on your own brand name?
Only when the clicks are incremental. If you rank first organically and no competitor is bidding against you, a share of that spend buys clicks you already had. Run a geo-split test, measure the organic replacement rate, and let the number decide instead of the convention.
What is brand keyword bidding?
Brand keyword bidding means running paid search ads on your own company or product names. Brands do it to control the top of their own results page, block competitors, and shape the message. It works. The open question is how much of that traffic you were getting anyway.
How do you know if your brand bidding is actually working?
Impression share and ROAS won't tell you — both look excellent on brand terms because the intent was already there. The honest answer comes from pausing brand ads in matched test markets and measuring what organic recovers. That's a brand search incrementality test.
If we stop bidding on brand, won't a competitor take the top spot?
Sometimes — and that's a real cost worth measuring rather than assuming. A geo split shows exactly what happens. Competitor conquesting appears as organic clicks failing to recover in the test markets, which pushes your replacement rate down and tells you to keep investing.
The bigger surface
Most of your brand SERP isn't yours
People searching your name aren't checking whether you exist. They're evaluating you, and most of what they find on that page was written by someone else.
“Brand SEO is an actual thing, and it's very neglected. A lot of customers are surprised when they actually Google themselves.”Jenna Kobryn · SEO & AEO Lead, Parallel Path
Yours
You write it and you own the placement.
- Top organic listings
- Sitelinks
- Your paid brand ads
- Your About and site content
Influenced
You feed it. Someone else decides.
- Knowledge Panel
- AI Overview inclusion
- People Also Ask answers
- Third-party profiles (G2, Crunchbase)
Out of your hands
It's there whether you like it or not.
- Competitor ads on your name
- Reddit and forum threads
- Review-site rankings
- Old or negative listings
Counting which of those you actually own is the first thing a brand search visibility audit does. The number is usually lower than the room expects, which is also why brand paid search feels indispensable long before anyone checks whether it is.
The method problem
The pause test isn't a test
Most brand incrementality "tests" work like this: switch brand campaigns off for a month, compare the traffic to the month before, draw a conclusion. That isn't an experiment. It's a before-and-after with no control.
What most people run
Month 1
Month 2
Four weeks apart — and these moved too:
- Seasonality
- Your promo calendar
- Competitor activity
- PR, news, and paid social
No control. The number can't be trusted.
What actually works
Test markets
Control markets
Same weeks. Same seasonality. Same promos. Same competitors.
The difference between them is the answer.
You'll find the first version published as a case study: brand paused through a quiet stretch, switched back on months later, the recovery read as proof the ads were working. Demand also returned in that window. So did the promo calendar. A control market holds all of it constant — which is the only reason the number means anything.
That's a difference-in-difference design, and it's the whole basis of a brand search incrementality test. Which leaves one question: inside those test markets, what exactly are you measuring?
The metric
How much organic actually comes back
The organic replacement rate is the share of paid brand clicks that organic recovers once you switch brand ads off. Divide the organic clicks you gain in the test markets by the paid clicks you gave up. The closer to zero, the more incremental your brand spend.
| Replacement rate | What it means | Action |
|---|---|---|
| 80–100% | Paid is mostly cannibalizing organic | Pause & reallocate |
| 40–80% | Partial incrementality | Reduce brand bids |
| 0–40% | Strong incremental traffic | Maximize investment |
Read these as decision bands, not statistical law. A 78% and an 82% are the same finding: a brand program that is mostly buying back traffic you already had. Which band the number lands in matters; where it sits inside the band does not. A result parked on a boundary is a reason to extend the test, not to split the difference.
What one read looks like
Illustrative — not a client result.
- Organic replacement
- 39%
- Incremental traffic
- 10,000
- Incremental conversions
- 650
- Incremental CPA
- $27
Verdict: brand paid search is incremental here. Keep investing, and hold the budget.
A number like that only means something if the test underneath it was built properly. Here's how to run one.
The protocol
How to run a brand search incrementality test
The design is a difference-in-difference: matched markets, one arm off, one arm on, the same weeks. It assumes Search Console is linked to Google Ads, because you need organic and paid on the same query to read the result.
Design
Before anything gets switched off.
- 1
Confirm there's enough volume
You need enough brand clicks in the test markets to detect a difference at all. Low-volume brands can still test — it just takes longer.
- 2
Match your markets
Pair test and control DMAs on brand volume, seasonality, and competitive pressure. Mismatched markets produce a number that means nothing.
Run
7 to 30 days, depending on volume.
- 3
Switch brand paid off in the test markets only
Control markets keep running exactly as they were. Change nothing else in either arm for the duration — no bid edits, no new promos.
- 4
Hold the window
Don't flip campaigns back on because week one looks alarming. Early noise is normal, and a test cut short is a test you paid for and can't use.
Read
Once the window closes.
- 5
Divide organic clicks gained by paid clicks lost
That's your replacement rate — measured in the test markets against what the control markets did over the same weeks.
- 6
Look past clicks
Total traffic, conversions, and incremental CPA are what tell you whether the clicks you gave up actually mattered to the business.
- 7
Decide: scale, reduce, or defend
The band your number lands in gives you the direction. The conversion data gives you the size of the move.
How long to hold the window
| Brand click volume | Hold the test for |
|---|---|
| High | 7–14 days |
| Medium | 14–21 days |
| Low | 21–30 days |
What you have at the end is a defensible number with a dollar figure attached to it — and a budget decision that doesn't rest on anyone's opinion about brand defense. Parallel Path designs and runs these as part of a brand search visibility audit.
What you don't have is a reason to run one on every account. Four conditions will hand you a number you can't trust.
Fit check
When to run it, and when not to
A brand incrementality test spends real time and real traffic. It earns that when a genuine budget question is on the table — and it hands you a number you can't use if the conditions are wrong.
Run it when
Any one of these is reason enough.
- Brand search is a large share of paid revenue
- You rank strongly organically for your own name
- Budget reallocation is on the table
- Competitors are bidding on your brand
- Someone suspects brand is inflating performance
- You need a true cost of acquisition
- A large awareness campaign is in flight
Don't run it when
Any one of these will hand you a number you can't trust.
-
Brand click volume is too low to read
The difference between the two arms disappears into week-to-week noise. Extend the window, or wait for a period with more volume.
-
You're inside a promo or a seasonal spike
Demand is moving for reasons that have nothing to do with your ads. A control market absorbs some of that, not all of it.
-
A competitor is actively conquesting your brand
You'll measure their campaign as much as your own. Either wait for it to stop, or go in knowing the result describes both of you.
-
Someone will flip the campaigns back on
A test rescued in week two costs you the spend and gives you nothing. Get the agreement to hold before you get the market list.
All of this assumes you already know which brand queries are worth testing. That answer doesn't come from the test — it comes from looking at paid and organic on the same query, before anything gets switched off.
Before the test
Find the redundant spend first
The test tells you whether brand paid is incremental. It doesn't tell you which queries to test. That comes from putting paid and organic side by side on the same search term — which most accounts have never done, because the two data sets live in different tools.
Link Search Console to Google Ads and you can read organic clicks and average position next to paid impressions, clicks and cost, query by query. Parallel Path calls that combined view the Search Query Intelligence Matrix. It's the first thing we build, and it usually reorders the priorities before a single campaign is touched.
What it surfaces
Illustrative brand queries — the shape of the output, not a client account.
| Brand query | Paid | Organic | The read |
|---|---|---|---|
| yourbrand | High spend | Position 1 | Redundant — test it |
| yourbrand reviews | Moderate spend | Position 1 | Redundant — test it |
| yourbrand pricing | High spend | Not ranking | Paid is earning it |
| yourbrand vs [competitor] | Nothing running | Position 7 | Gap — nobody owns it |
Notice that only two of those four rows point at cutting spend. The matrix finds gaps as often as it finds redundancy — queries where paid is doing work organic can't, and queries where neither channel shows up at all. A diagnostic that only ever recommends spending less isn't a diagnostic. It's a sales pitch with a table attached.
This is the foundation the rest of it sits on, and it's the same view that drives a holistic search strategy across every intent stage, not just brand.
The reallocation
Where the money goes instead
If the test says brand paid is mostly buying back traffic you already had, the recommendation isn't a smaller budget. It's the same budget aimed at queries where paid creates volume rather than recapturing it.
“You spend X every month, and it adds up to a pretty hefty annual investment — but you're ranking strongly organically. Run the test and find out whether that's actually the case. Then take the savings and put them against demand-generating queries organic is never going to fill.”Billy McCall · Paid Search Lead, Parallel Path
Three places that money does work it can't do on your brand name:
Consideration queries
Comparison and alternative searches where a competitor is already bidding and organic takes quarters to move.
Non-brand transactional
Bottom-funnel terms where paid controls placement and the first organic result sits well below the fold.
Research-stage testing
A fast read on which topics and messages actually convert, before SEO spends two quarters building for them.
Same budget. Different job. The test is what gives you permission to move it.
Our own numbers
We rank. Nobody clicks.
One Parallel Path blog post, branded-keyword queries, Google Search Console, trailing three months.
Position 27.9 is the reason, and we're not going to dress that up. The post is old, it sits on page three, and it's on the list to be retired. Zero clicks at that position is exactly what should happen.
What's useful is the shape of it. Four thousand impressions on one end, nothing at all on the other. The impressions told us we showed up. They told us nothing about whether it mattered.
Now look at how brand paid search usually gets defended: impression share. Ninety-plus percent on your own name, a strong top-of-page rate, a green cell in the monthly deck. That is the same class of metric. It measures appearance, not consequence — and it will look excellent on brand terms no matter how incremental the spend is, because the intent was already there before you bid on it.
The organic replacement rate is made of different material: clicks you gave up, clicks you got back. Both of those are things that actually happened. That's why it can settle a budget question and impression share can't.
Impression share tells you that you showed up. It has never once told you that you were needed.
Everything else
The questions that come up next
Can you bid on trademark keywords?
Yes. Google's advertising policy does not restrict using trademarks as keywords, including a competitor's. What it restricts is trademark use in ad text, and trademark owners can file complaints against specific advertisers on that basis. Local trademark law still applies and varies by country — worth counsel before conquesting internationally.
How long does a brand incrementality test take?
One to four weeks live, plus design before and analysis after. High-volume brands can read a result in 7 to 14 days; low-volume brands need 21 to 30. The constraint is click volume, not calendar. More brand traffic means a cleaner read sooner.
What counts as a good organic replacement rate?
Lower is better. Below 40% means brand paid is bringing traffic organic wouldn't have captured, and the spend is doing real work. Above 80% means it's largely recapturing clicks you already had. In between, reduce bids rather than pausing outright — partial incrementality is still incrementality.
How is this different from a brand lift study?
A brand lift study measures whether advertising moved awareness, recall or consideration, usually through survey panels. An incrementality test measures whether the ads produced traffic and conversions that wouldn't have happened anyway. One asks whether people remember you. The other asks whether you needed to pay.
Won't pausing brand campaigns hurt Quality Score or impression share?
Impression share drops in the test markets while ads are off — that's the test working, not a side effect. Quality Score is built on keyword-level history and generally rebuilds once campaigns resume, and brand terms rebuild fastest because relevance and click-through are already strong. Control markets run untouched throughout.
Does this change if there's an AI Overview on our brand term?
It changes what organic recovery looks like. If an AI Overview answers the query outright, some of the traffic that used to return as an organic click now doesn't return at all. That pushes your replacement rate down — and can make brand paid look more incremental today than the same test would have two years ago.
Should we bid on competitors' brand names?
Different question, different math. Conquesting buys clicks you were never going to earn organically, so incrementality is usually high by definition. The real questions there are cost and conversion rate, not cannibalization. Test conquesting on efficiency. Test your own brand on replacement.
What it takes to actually run one
A test is the third step, not the first. You need a baseline before you can measure a change, and a reason to believe a particular set of queries is worth testing before you switch anything off.
-
Start with an audit
Establish the baseline, map what you own on your own brand SERP, and find the queries where paid and organic are both already strong.
-
Then design the test
Feasibility against your click volume, matched test and control markets, hold window, and the measurement plan you'll read it with.
-
Then run and read it
Execution and analysis are scoped separately, because the answer only matters if someone acts on it.
Parallel Path builds all three as part of a brand search visibility audit — and will tell you when your numbers say not to bother.
Nobody has to win this argument
Brand search has been settled by conviction for fifteen years: protect the results page, block the competitors, own your name. Some of that is true for your account and some of it isn't. The only way to know which is to run the test and read the number it gives you.
The bigger picture
Building a holistic search strategyBrand is one intent stage out of four. The same logic runs across the rest of them: channel roles, shared data, measured overlap.
Keep reading