Most line extensions are too small to lift the brand they came from
We screened 45 extensions across the same aisle, a new aisle and licensed partners. Four were big enough to test for a halo, none of those showed a clean one, and the place the parent and the extension do meet is the AI answer.
A line extension usually arrives at the approval meeting with two cases attached. One is the product’s own: what it sells, at what margin. The other is softer: the new product will remind people of the brand they already buy, and lift it. That second case rarely gets a number, because nobody expects it to need one.
So we tried to give it one. We took 45 line extensions and sorted them three ways: the same aisle as the parent (Twix Cookie Dough, Bud Light Seltzer), a new aisle (Oreo Cakesters, Liquid Death iced tea), and licensed to a partner (Taco Bell Cravings Kits, Starbucks Creamer). Then we measured three things the halo would have to show up in: what people search for, what AI assistants say, and what reviewers write. Every prediction was written down before the data was pulled.
Most extensions are small next to the parent
An extension can only send people back to its parent if people notice it. We set a gate before looking: in its launch month or the month after, the extension had to draw at least a fifth of the parent’s usual monthly searches. Below that, any move in the parent is more likely to be weather, a holiday or an ad than the new product.
Line extensions that drew at least 20% of their parent brand’s searches at launch, the minimum for a halo to be measurable at all. None of the 11 licensed ones did.
Launch months were read from each extension’s own searches and checked against a dated public announcement. Units whose dates did not match were dropped.
Even the ones picked because they looked big mostly failed. Skittles Gummies drew 1.7% of Skittles’ searches in its launch month. Bud Light Seltzer drew 2.8% of Bud Light’s. Starbucks Creamer drew less than 1% of Starbucks’. The four that passed were Oreo Cakesters, Liquid Death iced tea, Oatly ice cream and Heinz Pickle Ketchup, the last on the strength of its announcement.
Three tests, and no clean lift
For three of the four we pulled Google Trends, which reports search interest without rounding it into steps, and compared the parent’s launch month with the same month in every other year. That second comparison matters. A drinks brand rises every spring and a cookie brand peaks every December, and a launch timed into either looks like lift.
Change against the three months before, averaged across two pulls of Google Trends, United States web search.
Oreo fell 9% in the Cakesters month, which is a normal January after the holidays. Liquid Death rose 51% when its iced tea launched, well above its typical March, but in March 2024, with no launch, it rose 49%. One launch-free year matching the launch is enough to stop us calling it lift. Heinz is the one launch-shaped move: up 17% in the month Pickle Ketchup was announced, the only November in seven years above 8%, and still under the bar.
Getting here took three attempts, and the first two failed on our own design. The first used a 20% bar on data Google rounds in steps of about 22%, so a real lift could read as zero. The second read spring peaks as lift. Both are published with the study. What survives is narrow: among extensions big enough to test, none produced a halo that a year without a launch didn’t also produce.
The assistant connects them, and picks the source
Search is where the halo didn’t show. The AI answer is where parent and extension are tied together without anyone asking. We asked ChatGPT about the product range of seven parent brands, twice.
It named the extension for six of seven brands in both runs. For owned extensions it described them from the owner’s own site every time: coca-colacompany.com, oldelpaso.com. For licensed ones it never cited the restaurant or the company making the product. The answers about Chick-fil-A sauces, Taco Bell kits and IHOP syrups came from retailers, scraped pages and an unofficial menu site. For Ghost, a lookalike domain ranked first in one answer. Both Coca-Cola answers listed Spiced as a recent launch; it has been discontinued.
Reviews point the same way for licensed products. Of ten reviews of Chick-fil-A sauces and Taco Bell kits, four said the product changed what the reviewer does with the restaurant, and three of those were about eating at home instead. Two said the kit replaces a Taco Bell visit. None said it sent them back.
- The extension will draw a fifth of the parent’s attention at launch, by your own forecast. Below that, nothing it does to the parent can be measured, so it can’t be in the case.
- You have checked the parent’s own seasonal pattern for the launch month, across several years, before reading any launch result.
- For a licensed product, you have counted the visits it replaces, not only the royalty it earns.
- You know what assistants say about the parent’s range today, and whose page they use to say it.
If an extension clears all four, a halo can be tested and, if it shows, counted. Most won’t clear the first, and for those the honest case is the product’s own: what it sells, at what margin, with no credit for the brand it came from.
- CEO Only 4 of 45 extensions were big enough for a halo to be tested, and none showed a clean one. Approve each extension on its own P&L and drop the halo as a justification.
- CFO Licensed products earn a royalty and, by four of ten reviews, replace some restaurant visits. Net the replaced visits against the royalty before the license is signed.
- CRO ChatGPT named the extension for 6 of 7 parent brands, so buyers asking about the parent already hear about it. Make the owner’s site the page that answer is built from.
- CPO Same-aisle reviews compared every product to the parent; new-aisle reviews compared them to competitors. Design a same-aisle extension to trade buyers up inside the brand, and a new-aisle one to win its aisle without help.
- Legal A lookalike domain was the first source cited for one brand, and licensed products were described from unofficial sites. Write answer monitoring and lookalike takedowns into licensing and trademark terms.
- Risk Assistants described a discontinued product as current. Check what assistants say about every live and retired extension each quarter.
Reconciled. The CEO and CFO seats pull toward cutting extensions that lean on a halo; the CRO and CPO seats pull toward keeping them, because buyers and assistants already connect them to the parent. One rule satisfies both: an extension clears its own return hurdle with no credit for parent lift, and the money that would have gone to proving a halo goes to the owner’s page that every assistant reads.
Run the gate before the approval meeting.
Take the extension’s forecast launch-month searches and divide by the parent’s average monthly searches over the last three months. If the ratio is under 20%, remove the halo line from the business case. Then ask two AI assistants about the parent’s range, three times each, and note whose pages they cite.
- What it costs
- An analyst’s afternoon and under a dollar of search and assistant data. The real cost is a business case that may lose its halo line before it reaches the committee.
- How you’ll know
- The ratio is a yes or no on the day. The assistant check is working when the owner’s own site is the first source for the extension in two of three answers.
Put a number on the extension before it gets a budget.
The CPG Decision Brief takes the six decisions a brand manager owns, extensions among them, and answers each from sourced, dated evidence with its limits stated. Or commission one built around your own portfolio.
See the CPG Decision Brief Or commission a custom brief →