Discussion starter: an SEO measurement question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.
- SEO Analytics
Has anyone paused branded paid search to measure how much organic absorbs? What happened, and would you repeat it?
Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.
1 reply
-
adminAdmin
Brand pause tests have been run and published, and the results point in both directions. eBay’s large field experiments, written up by Blake, Nosko and Tadelis, found that brand-keyword ads had no measurable short-term benefits, which the authors present as an extreme case. A 2012 Google meta-analysis of 390 ad pause studies found that, on average, 50% of the ad clicks that occurred alongside a top-ranked organic result were incremental, compared with 100% when there was no associated organic result.
Both studies are more than a decade old, and the gap between them is the useful part. The variable I would look at first is not your organic ranking, which is probably already first for your own name, but who takes the top of the page when your ad disappears: a competitor, a reseller, an affiliate, a marketplace, or nobody. I would repeat the test, but split by query type, protected from leakage into other campaigns, and judged on incremental margin rather than on how many clicks organic absorbed.
Why these two results disagree
eBay is a favourable case for pausing: a household name, where people searching the brand already know where they want to go. The Google figure averages 390 studies across many advertisers, and it comes from the company that sells the ads. That doesn’t make it wrong, because pause studies are a reasonable method, but it is an average across businesses unlike yours. Results pages have also changed a great deal since both were published. Treat the two findings as bookends, not as a forecast for your account.
What organic can absorb, and what it cannot
Think of a brand ad as a member of staff standing at your front door. On a quiet street, customers walk in anyway and the greeter adds little. If a competitor sets up a stall on your doorstep, the greeter is what stops some customers wandering over to it. A pause test mostly measures whether that stall exists. In search terms, the stall is any other advertiser eligible to show on your brand terms, plus any organic result that competes for the click.
When the ad is paused, the clicks it used to take go to one of five places:
- Your organic listing. This is the absorption you are hoping for.
- A competitor bidding on your brand. That is a lost customer, not just a lost click.
- A reseller, marketplace or affiliate. They may still sell your product, but at a commission or a lower margin, which can cost more than the click did.
- Another organic result, such as a review site, coupon site or marketplace listing.
- Nowhere. The searcher gives up, or returns later through another channel.
The mix shifts with query type. Pure navigational searches (the brand name alone, or the brand plus “login”) are likelier to be absorbed in full when nobody else bids. Brand plus a modifier, such as “discount code”, “reviews” or “vs [competitor]”, attracts coupon sites, review sites and competitors, so less of it comes back through organic. Treat these as separate tests, because a blended result can hide a segment that should keep its ads.
Before pausing, look at your brand results page as a searcher would. If a reseller outranks your own product pages, the sitelinks point to the wrong sections, or a page of negative reviews sits in second place, the ad has been covering an organic problem. Fix that first, or the test will measure the organic problem rather than the ad.
How I would set up the test
Design How it works Strength Weakness Full pause, before and after Pause everywhere and compare with the previous period Simple and fast Seasonality, promotions and algorithm updates contaminate the comparison Geographic split Pause in some regions and keep ads running in matched regions Controls for seasonality and national campaigns Needs enough volume per region; location targeting and travel blur the boundaries Switchback Alternate on and off periods, for example week by week Handles seasonality without needing regions Carryover: competitors and customers can react across periods - Choose the segment. Split brand queries into navigational and modified groups, and decide which one you are testing.
- Stop leakage. Performance Max, broad match and other automated campaign types can match brand searches, so a paused brand campaign may simply move its traffic elsewhere in the account. Apply brand exclusions or negatives to every other campaign for the test period, and check search terms daily for the first week.
- Check affiliate and reseller terms. If affiliates are allowed to bid on your brand, pausing can shift clicks into commissions.
- Watch the results page yourself. Once you stop bidding, Auction insights no longer shows who else bids on those terms, because it only reports on auctions you take part in. Record which ads and results appear in each test region before and during the test.
- Judge on total outcomes. The primary metric is total conversions and revenue in the test regions or periods, across all channels, compared with the control. Pausing moves conversions between channels, and channel attribution moves with them, so a channel-level view will mislead.
- Track the mechanism with secondary metrics. Use organic brand clicks from Search Console’s branded queries filter (spot-check what it classifies as branded), paid brand clicks and cost, direct traffic, and the share of new customers. Search Console filters by country, not by state or city, so a geographic split within one country has to be read through GA4 and sales data instead.
- Size it before you start. Decide the smallest drop in brand conversions that would change your decision, and check that your volume can detect it within the test period. A low-volume test can end inconclusive, and an inconclusive result is easy to misread as “no effect”.
- Avoid peak periods. Running the test through end-of-financial-year sales or Black Friday mixes promotional effects into the result.
On duration, run each condition across several full weeks, because brand demand varies by day of week, and long enough to cover the typical time from first visit to purchase or lead. A retailer whose customers buy within days can read a switchback in a few weeks per arm. A B2B business with a long sales cycle will see clicks and form fills well before pipeline, so decide in advance which of those the decision rests on.
The maths that decides it
Three numbers matter:
Absorption rate: organic brand clicks during the pause, minus organic brand clicks with ads on, divided by paid brand clicks with ads on.
Lost clicks: paid brand clicks multiplied by (1 minus the absorption rate).
Decision: keep the ads if lost clicks × conversion rate × margin per conversion is larger than the brand spend you would save. If you measured total conversions directly, compare the lost margin with the saved spend and skip the click step.
Hypothetical example:
A retailer’s brand campaign gets 10,000 paid clicks a month at an average of $0.30, so it costs $3,000. Organic brand clicks run at 30,000 a month. In the paused regions, organic brand clicks rise by enough to recover 75% of the paid clicks, and the other 2,500 go to a competitor’s ad and a coupon site. If brand visitors convert at 5% and each order earns $40 in margin, those lost clicks cost about $5,000 a month in margin against $3,000 saved. A 75% absorption rate sounds like a case for pausing, yet the ads more than pay for themselves. With no competitor on the page and 97% absorption, the same maths gives about $600 of lost margin against $3,000 saved, and pausing wins.
The example also shows why “organic absorbed most of it” is not a verdict on its own. When brand traffic converts well, a small share of lost clicks can outweigh a cheap campaign.
Would I repeat it?
Yes. It is cheap to run and easy to reverse, and the answer expires. A test run while nobody bid on your brand says little about next quarter, when a competitor or a new reseller may start. After a first test, I would:
- keep ads off for segments where absorption was high and the margin maths favoured pausing;
- keep ads on for modified brand queries where competitors, coupon sites or review sites took clicks;
- check the brand results page weekly and turn ads back on when a competitor appears;
- rerun a short switchback when competition, pricing or the affiliate program changes.
What would change the decision: if total conversions in the paused regions fell by more margin than the spend saved, or competitor ads appeared within days of the pause, I would switch brand ads back on and treat them as insurance. Near-complete absorption with no competitor activity across several weeks is the case for keeping them off.
If this were my account, I would stop asking whether brand ads are incremental in general. I would ask which brand queries, in which regions, face which competitors, and what a lost click costs in margin. Then I would test those segments separately, decide on profit, and schedule the next test before the current one ends.
Need help with this on your own site?
Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.
Add a reply Cancel reply
Replies are open to approved members. Log in or apply to join.