Spec work · nobody asked for it
Lyko Group · Sweden

The Cost of Being Well Known

Lyko told the market that strong brand awareness had let it lower marketing spend “without effect on performance.” I think the effect is already visible. It is just not where that sentence was looking.

Brand
Lyko Group
Market
Sweden
Discipline
Brand demand & measurement
Written
September 2026
This is unsolicited spec work and it is not affiliated with Lyko Group. Nobody commissioned it, nobody at Lyko was consulted, and every figure in it comes from the public sources listed at the end. There are no results, because the work was never run. What stands in their place is a hypothesis and a plan for measuring it.

The business problem

Lyko Group is the Nordics’ largest beauty e commerce specialist: around 65,000 products from over 1,000 brands, e commerce in eight European markets, 36 stores in Sweden, Norway and Finland, and about 1,100 employees. It passed SEK 4bn in rolling twelve month net sales in mid-2026.

The profitability story is genuinely good. The growth story underneath it is not.

Q2 2026 delivered net sales of SEK 1,021m, up 8.6%, and EBIT of SEK 54m at a 5.3% margin, the company’s best second quarter ever, on the back of a SEK 100m cost reduction programme and newly automated warehouse capacity. Over the same half, growth fell to +2.0%, against +11% for the full year 2025.

Figure 1
Quarterly net sales growth, year on year
Lyko Group, Q2 2025 to Q2 2026. Hover or focus a column for its value.
15 10 5 0 −5 % +6.8% +8.5% +17.0% −4.8% +8.6% +17.0% −4.8% Q2 ’25 Q3 ’25 Q4 ’25 Q1 ’26 Q2 ’26

Source: Lyko Group interim reports, Q2 2025 through Q2 2026. Full figures in the table below.

The same figures, as a table
PeriodNet salesGrowth
Q2 2025SEK 939m+6.8%
Q3 2025SEK 827m+8.5%
Q4 2025SEK 1,278m+17.0%
Q1 2026SEK 875m−4.8%
Q2 2026SEK 1,021m+8.6%
H1 2026SEK 1,895m+2.0%
FY 2025SEK 3,963m+11%

Two things happened around it

One, the competitive structure changed. KICKS is now part of Matas Group, and the two describe the combination as the leading beauty omnichannel group in the Nordics, in a market they size at over DKK 65bn. It is a strong and well resourced group, and the category got harder for everyone in it.

Two, Lyko lowered its marketing spend.

“We have built strong brand awareness in Sweden and Norway, which allowed us to lower marketing spend during the period without effect on performance.” Lyko Group, Q2 2026 report, 17 July 2026
kicks.se
1.55M
estimated monthly visits, July 2026
lyko.com
1.02M
estimated monthly visits in the same month, in a category Lyko led on net sales in 2025

Source: Semrush traffic estimates, Swedish beauty & cosmetics, July 2026. Modelled, not measured. See the limitations.

The insight

“Without effect on performance” is a claim about a quarter, not about a brand, and Lyko is already holding the evidence that reads the other way.

Awareness is a stock, not a flow. Stop topping it up and it does not collapse; it stops being replenished, and the first place that shows is never revenue. It shows at the top of the funnel: who people think of first, how many arrive without being bought, how many come back without being retargeted. Revenue moves last, which is exactly why cutting brand spend looks free for two or three quarters.

Both readings sit in Lyko’s own numbers. EBIT is the best it has ever been in a second quarter. Growth across the half year is 2.0% against 11% the prior year. The sentence in the report picks the first and calls it performance.

Lyko’s awareness was bought, not built, and bought awareness decays the moment you stop paying.

The brand memory a performance led retailer accumulates is transactional. People remember you at the moment they were already shopping, because you were the result they were served. It is attached to the transaction, not the brand, so when the spend stops there is nothing holding it in place. Awareness that survives a cut is attached to something distinctive the money was only amplifying.

Lyko has exactly such a thing, and files it under About.

Lyko is not an e commerce company that sells hair products. It is a family of hairdressers, three generations deep, that happens to run the Nordics’ largest beauty retailer.

1952Frans Lyko opens a men’s hair salon in Vansbro, Dalarna.
1971His son Stefan opens Hår & Skägg in nearby Dala-Järna.
2003Stefan registers lyko.se, on 5 May.
2008Rickard Lyko founds Lyko Group with Rita, Stefan and Erika Lyko.
2026SEK 4bn, eight countries. Head office still partly in Vansbro.

It took three generations to make, which is exactly why it cannot be bought, copied or outspent by anyone. It is also the asset Lyko is not currently using.

A judgement call, stated rather than buried

Frans Lyko survived nearly seven years in Sachsenhausen, where his trade as a hairdresser is credited with keeping him alive. It is published on Lyko’s own history page and it is the most affecting thing in the company’s story.

It is not campaign material, and this strategy does not use it. It is the family’s history to tell, in their words and at their choosing, not a hook for an outside candidate’s speculative case. The usable asset is the craft lineage: three generations of people who do hair for a living. Where a story like that belongs is a decision for the family, not for a strategy deck.

The strategy

Target. Two occasions rather than a demographic. The reorder. They know what they want and go wherever is top of mind, which is where a traffic deficit becomes a revenue deficit. The wrong purchase. They don’t know what they need, buy something unsuitable, don’t return it, and quietly never reorder. That is the hidden cost of 65,000 products with no guidance layer.

Positioning. From the widest range online to the hairdressers who built the shop. Range is a claim any funded competitor can match. Craft lineage is not.

Message. Expertise with a provenance. Not “we are old”, but “we have been looking at people’s hair and telling them what it needs since 1952, and we still do.”

The constraint, which is the interesting part of the brief. Lyko’s narrative currently rests on the cost story. Any answer that simply restores the marketing budget undoes the thing the market is paying for. So the campaign has to build durable memory out of assets Lyko already owns and already pays for, and that points straight at 36 stores and ~1,100 employees, a large share of them trained hairdressers. It is the biggest owned media network in Swedish beauty, currently treated as a retail cost line.

The idea: Håranalysen

The Hair Consultation. Put the hairdresser back at the front of the business.

It is less a campaign than a service that quietly does a campaign’s work. Instead of landing on a search box in front of 65,000 products, you land on a short guided consultation: a few questions about your hair, and then a named Lyko hairdresser’s answer. The same consultation exists in the 36 stores, free and bookable. The same people appear on the product pages, with their salon and their years in the trade attached to what they recommend.

What makes it work is that it goes straight back into Lyko’s roots. The company starts in 1952 with a men’s salon in Vansbro, and underneath the e commerce it is still a family of hairdressers. Håranalysen takes that lineage out of the history page and puts it back at the front door, where a customer meets it instead of reading about it. Everything in the table below is built out of something Lyko already owns and already pays for.

AssetChannelWhat it does
The diagnostic. A short guided consultation replacing “search 65,000 products” as the front doorOwned: site and appTurns range from paralysis into service; captures data on what people actually need
Named recommendations. Every product page carries a real, named Lyko hairdresser’s verdict, with their salon and years in the tradeOwned: product pagesAttaches memory to people, not price. Uncopiable without salon staff
In store consultations, booked and freeOwned: 36 storesConverts a cost line into an acquisition and brand channel
Sedan 1952, a documentary series of Lyko hairdressers diagnosing real hair, shot in Vansbro and in storeOwned social, earnedThe distinctive brand layer. Craft, not catalogue
The reorder note. Reorder emails carry the consultant’s own follow up rather than a discount codeOwned: CRMDefends the reorder occasion without discounting margin
A small, consistent paid brand layer, deliberately not always on performancePaid, minority of budgetReach beyond the existing base

Sedan 1952 har vi tittat på folks hår och sagt vad det behöver.

“Since 1952 we’ve looked at people’s hair and told them what it needs. Now we do it online too.”

Hypothesis

Spec work has no results. Inventing them is disqualifying if anyone checks. What follows is what I would have measured and what I would have committed to.

The hypothesis. Lyko’s shortfall in unpaid traffic is a brand demand problem rather than a channel problem, and putting the paid spend back will not fix it. Brand activity built on guidance and on assets Lyko already owns will move unprompted consideration inside three quarters, and consideration will move the reorder rate before it moves revenue.

MeasureWhy this oneTarget
Unprompted brand consideration, SwedenThe stock the cut was drawing down, and the leading indicator nobody is currently reporting+X pp / 3 qtrs
Direct and organic share of sessions, against the categoryDemand you do not pay for, which is where the gap opened in 2026close by Y%
Reorder rate at 90 days, consulted against holdoutThe wrong purchase thesis, tested directly+Z pp
Consultation through to first orderWhether the service earns its operating cost> channel avg
EBIT margin (guardrail, not a target)The premise is rebuilding demand without rebuilding cost. If margin falls, the strategy failed on its own termsnot below 5%

The measurement design matters more than the targets

Comparing people who chose a consultation against people who didn’t measures intent, not the consultation. Anyone who opts in was already more engaged. Without a control this produces a large, flattering, meaningless number, which is the most common way an honest figure turns into a dishonest claim.

So, a matched market holdout. Hold three comparable Swedish store catchments out of the in store and CRM layers entirely, and read the difference. It costs some of the upside and it is the only version whose numbers would survive being checked.

Targets are left as X, Y and Z deliberately. Setting them needs Lyko’s baseline consideration tracker and CRM cohorts, which are not public. Numbers invented to fill the gaps would make the whole document worthless.

What this case cannot claim

The evidence has a ceiling, and it is better said here than found later.

  1. The traffic figures are modelled, not measured. Semrush estimates traffic rather than observing it, so read those two numbers as direction and nothing firmer.
  2. Nothing in here establishes cause. The traffic gap may predate the lower marketing spend entirely, and the first quarter of 2026 has competing explanations in the comparison base, the warehouse transition and the calendar. I have assumed an answer to neither. Both are the first things I would put real data against.
  3. All of it is built from public sources. No media mix, no consideration tracker, no CRM cohorts, no category margin, and one date I could not confirm in a primary filing. A real version of this document would open by asking for those.

What I would do differently

I built this with Claude, and it is worth saying plainly how: I set the question, ran the research and made every judgement in it, and used AI as the reader who keeps asking what the evidence is. What I would change is the order I worked in. I started with the financials and arrived at the brand argument second, when the interesting question was always what kind of awareness Lyko had actually built. Next time I begin there, and let the numbers describe the symptom, which is all they were ever going to do.

Sources

Every claim above traces to one of these. Read 8 September 2026.

Primary: Lyko’s own reporting
Secondary