Branded search defense for SaaS

Search your own brand name on a phone, logged out, on a network that is not your office WiFi. Count the ads sitting above your organic listing. That number, not a benchmark and not the CFO’s instinct, decides whether brand defense earns its budget.

The argument runs identically in every SaaS company. Growth points at a $9 cost per acquisition on brand terms and calls it the best line in the account. Finance points at the same line and says you are renting traffic you already own. Neither side can win the argument with the data they have, because a conversion report tells you what happened while ads were running and nothing at all about what would have happened without them.

There is a cheap way to settle this. Four weeks, a geo split, a spreadsheet.

What brand defense actually protects against

Three things, and only one of them is a competitor. In the order they most often bite SaaS companies: your own stale pages, affiliate and review-site arbitrage, then rival conquesting.

Your own pages come first because nobody talks about it and it takes thirty seconds to check. Type your brand name and look at what owns the top organic slot. For a surprising number of SaaS companies it is a 2024 post about a feature that shipped differently, a status page, a changelog entry, or a G2 profile carrying a 3.9 average and a competitor comparison widget. A paid ad is the only mechanism that reliably puts the page you chose, with the sitelinks you chose, above all of it. That is message control rather than click defense, and it earns its money during a pricing change, a rebrand or an outage even if it earns nothing the rest of the year.

Affiliate and review-site arbitrage is the quiet leak. An affiliate bids on your brand name, takes a click you would have had for free, passes it to your own site, then invoices you a commission on the signup. Most mature SaaS affiliate programs ban brand bidding in their terms, and enforcement is manual: somebody has to check the ads from three or four geographies every week, because affiliates geo-target around your head office. Review marketplaces run the legitimate version of the same trade. A Capterra or G2 category page bidding on your name is selling your buyer a shortlist with four of your rivals on it.

Rival conquesting is the one everyone argues about, and the only one of the three where defense changes click volume in a way you can measure. Whether it matters to you depends entirely on your category. Project management, CRM and HR software are permanent bidding wars. Search Monday.com or ClickUp and you will usually find each other’s ads. Infrastructure and developer tools are far quieter, and plenty of SaaS companies with a five-year-old category to themselves are paying to defend an auction nobody else has entered.

How much of a brand click is actually incremental

Most of it is not, until a competitor shows up. The most useful finding in the published research is that the incremental value of a brand ad is not a property of your brand at all. It is a property of the auction you happen to be in this month.

The reference experiment belongs to eBay. Blake, Nosko and Tadelis ran a large scale pause of eBay’s brand keyword advertising and published the result in Econometrica in 2015: returns to brand search were statistically indistinguishable from zero, because almost every click the ads bought would have arrived through the organic listing anyway. Google’s own research, Incremental Clicks Impact of Search Advertising from 2011, reports a median incrementality near 89 percent across search ads generally. No independent replication has come close to that figure. Treat it as the vendor’s number rather than the field’s.

The work that reconciles the two matters more than either. Simonov, Nosko and Rao, writing in Marketing Science in 2018, studied brand keyword auctions across many advertisers and found that the loss from pausing depends almost entirely on competitor presence. With nobody else bidding and a strong organic result, the clicks come back. With rivals occupying the slots above your listing, a real share of them go to the rival instead of down the page to you.

Those bands come from our own account set and should be treated as priors, not findings. The point of the table is the shape, not the decimals: brand incrementality is a step function keyed to how many other advertisers are in the auction, and the step happens around the second rival.

How to size the brand budget without picking a number out of the air

Derive it from volume and impression share. A brand budget set as a flat dollar figure will either starve during a launch spike or overspend through a quiet quarter, and in both cases nobody notices for months.

The arithmetic is four inputs. Monthly brand search volume, multiplied by target impression share, multiplied by brand ad click through rate, multiplied by brand cost per click. Brand ad CTR in a defended SaaS account usually sits between 20 and 35 percent, and lands at the lower end when your own organic listing is directly underneath, because some share of people scroll past the ad on purpose.

Target impression share deserves its own argument. Aim at 85 to 90 percent, not 100. The final slice of impression share is bought on the loosest query variants, misspellings and semantic matches, which is exactly where Google’s broad matching drags you into paying for queries that were never about you. That last 10 percent routinely costs as much as the previous 40.

Two sanity checks belong on top of that number. If brand is running above 25 percent of your total search spend, the problem is almost never brand, it is that your non-brand program is too small to be visible next to it, and the fix sits in SaaS PPC budget allocation rather than in the brand campaign. And if your brand CPC has doubled in a quarter without your bids changing, somebody new has entered the auction. Go and find out who, because that changes the incrementality answer completely.

The geo holdout that settles the argument in four weeks

Split your markets into two matched halves, run the brand campaign in one half, exclude the other half from the campaign’s location targeting, and compare total brand driven conversions across both. Paid plus organic, not paid alone. That last part is where most in-house tests fall apart.

Google Ads drafts and experiments will offer you a cookie-based 50/50 split, and for brand terms that split is contaminated before it starts. Brand searchers come back. They come back on a laptop after searching on a phone, three days later, from a different network, and the experiment assigns them fresh. Geography does not do that.

Now the part that gets skipped. Take a company with 4,000 brand searches a month in the United States, split into two matched halves of roughly 2,000 each. Assume brand intent traffic converts to trial or demo at 6 percent, so each arm produces about 120 conversions over four weeks. At that sample you are powered to detect a swing of roughly 35 percent, and nothing meaningfully smaller. Run eight weeks and you get down to about 25 percent.

That looks like a failed test design. It is not. The question in front of you was never whether brand ads are 12 percent incremental or 18 percent incremental. It is whether switching them off costs you a chunk of pipeline or close to nothing, and a test powered to catch a 35 percent collapse answers exactly that question for the price of four weeks of forgone clicks in half your territory.

When a time split is the only option you have

Single market companies cannot geo split, and pretending otherwise produces two halves of one country that share a media schedule, a PR cycle and a sales team. The alternative is a time split, and it is a genuinely weaker instrument.

Run two week blocks, alternating on and off, repeated at least three times for a twelve week test. Two week blocks matter because a one week alternation puts every weekday pattern in the same phase, and SaaS brand search volume is heavily weekday loaded. Three cycles matter because anything less cannot separate the treatment from a single bad month.

The confounds you cannot remove are seasonality, a competitor’s campaign starting mid-test, a funding announcement, a podcast mention, and your own content program shipping something that ranks. Keep a dated log of all five during the run and read the result with them in front of you. A time split that comes back flat is fairly convincing. A time split that comes back showing a large loss deserves a second cycle before you act on it.

Reading the result: three outcomes and what each one means

Flat, small loss, large loss. Each has a different correct response, and only one of them means switch the campaign off permanently.

A flat result, meaning the holdout half converted within a few percent of the treatment half, says your paid brand clicks were cannibalising organic ones. Cut impression share to somewhere between 20 and 40 percent rather than going to zero, keep exact match only, and set a weekly SERP check. Holding a small presence keeps quality score history alive and gives you a running measurement of rival entry, which is worth more than the handful of dollars it saves.

A small loss, in the range of 10 to 25 percent of brand conversions, is the most common result in a category with one intermittent rival. Keep defending, but stop paying for the top of the impression share curve. Push target impression share down from 95 to around 70 and watch whether the loss reappears.

A large loss says your organic listing is not holding the click on its own, and there are only two reasons for that: rivals are taking it, or your own organic result is weak. Check which before you conclude anything. If a G2 category page or a stale blog post is outranking your homepage for your own brand name, the paid budget is treating a symptom, and the cheaper fix is an afternoon of on-site work.

When pausing brand ads is genuinely safe

All of the following, simultaneously, and verified rather than assumed. Any single one missing and the answer flips back to defend.

That fourth item costs some companies more than all the others combined. Brands named after ordinary words pay a permanent tax on this. Monday, Linear, Ramp, Clay, Loom, Segment, Drift and Gong all have to buy their way past dictionary intent, and their brand campaigns need a negative keyword list long enough to be its own project. If your brand name is a common noun, you are not really running a brand campaign, you are running a disambiguation campaign, and the pause decision does not apply cleanly to you.

What the brand campaign should look like if you keep it

Exact match only in the core campaign, a separate campaign for brand plus modifier queries, and a negative list that removes people who are already customers. Most brand campaigns are built once and never touched again, which is how they end up paying $6 a click for people trying to log in.

Strip these out with negatives: login, sign in, status, down, outage, careers, jobs, salary, support, refund, cancel, and the name of your own help centre. Those searches are support traffic and you pay real money for them. Keep brand plus pricing, brand plus alternatives, brand plus versus, and brand plus review in their own campaign with different copy, because the intent is different and so is the right landing page. The starter set for all of this lives in the SaaS negative keyword list.

The ad copy is where most teams waste the slot. A brand headline that repeats your own title tag adds nothing, because the person already knows who you are. Use the space to pre-answer whatever objection makes people bounce: transparent pricing, the integration they are checking for, SOC 2, a migration path off the tool they are currently paying for. Sitelinks should point at pricing, security, the migration guide and your strongest comparison page, which is also the page that has to catch anyone arriving from a rival’s conquest ad. There are worked examples of that in the SaaS ad copy swipe file and live teardowns in SaaS Google Ads teardowns.

One more structural decision. Brand and conquest campaigns should both run on manual CPC or a hard bid cap, not on target CPA. Brand conversion rates are so far above the account average that automated bidding will happily raise your brand bids to the point where you are paying $14 to defend a click you would have had for $2. That specific trap, along with the rest of the greatest hits, is covered in the PPC mistakes that quietly drain SaaS accounts.

What to do in the next two weeks

Run the sixty second audit from the top of this page and count the rivals. If the count is zero, build the geo holdout and get the four weeks started, because you are probably spending real money on nothing and you will never get a cheaper answer. If the count is two or more, skip the holdout for now and put the effort into the comparison page that has to receive both your defensive clicks and their conquest clicks.

Then price the decision properly rather than arguing about it. Work out what a brand click is worth to you using the SaaS max CPC calculator, compare your brand CPC and share of spend against the SaaS PPC benchmarks, and read the other side of the trade in competitor brand bidding for SaaS, because the people bidding on you are running the same arithmetic in reverse.

The position worth holding is narrow. Nobody should defend brand terms on faith, and nobody should cut them on principle either. A four week geo holdout costs less than a year of arguing about it, and the result is only valid until the next competitor enters the auction, which is the real reason to keep a small brand presence running and a weekly alert switched on. The rest of the account structure sits in the SaaS PPC and paid ads hub.