Uses Google shopping “directly” instead of using a CSS like Kelkoo for 20% cheaper clicks.

Make an agreement with Kelkoo and then do some changes in merchant center. Other than that no other changes need to be made and the “discount” is given automatically, no changes need to be done in either feeds or the Google shopping campaign.

Bidding on brand

In general, when bidding on brand, you will pay for clicks that would have been free. So it 100% leads to wasted money, the only question is if you “save” some sales from competitors bidding on your brand.

In general I advice against it and just accept that you might lose a few sales to competitors.

If bidding on brand name, having a low budget is reasonable, 100 kr is the daily budget now, which is fine.

But the bidding strategy should IMO be manual and then put so low that you do not use up the budget.

For example, now the avg cost is 1.33 kr per click, last 30 days. I think if one puts the max bid around 0.5-0.8 kr per click on “kashmina” phrase match, then you will get more clicks for lower cost. Over time this will definitely add up, better protecting the brand, at a lower cost.

Let’s say that bid is set to 0.5 kr, then if just 50 kr of the 100 kr budget is used. Then it would mean you still get more clicks on 50 kr than the 100 kr used if avg cost is 1.33 kr per click.

Performance max

In general, I do not trust Google. It is convenient, and might be good since it automatically includes placements where you normally would not “put in the time” to test.

Shopping campaign

The Shopping campaign sets a ROAS of 500%, I think that is a reasonable target, we see that the budget is often not used up. Which I think is a good thing, that means you set so high ROAS that Google “decides not to waste your money”. Too high bidding in manual max CPC or too low target ROAS will just lead to Google waste your money on high cost clicks.

One of the reasons to use manual bidding is to avoid absurdly high CPCs. Last 30 days over 30 clicks had higher than 20 kr CPC under performance max, and similar for the normal shopping campaign with target ROAS 500%. In both campaigns you have 1 click with 50 kr CPC, which definitely is not good.

The counter-argument will be that “Google is letting you pay more for clicks that are higher quality on high value items”, I generally think these excuses are snake oil.

So what I recommend is trying to do manual shopping and then do max CPC per product type, using a higher max CPC for the product that are more profitable.

SEO

SEO is a long term game, let’s say that it will be relevant to sell to Sweden/Denmark in 2 years. It will be very beneficial to start setting up kashmina.se and kashmina.dk now, and just use translate & adapt app which has free auto-translate of 2 languages to translate the sites.

I would also consider buying the other country level domains kashmina.at, kashmina.be etc. for brand protection and future proofing. At least for the main countries in Europe.

Meta titles

Good meta titles, the title that is shown in search results is important for CTR.

Example of possible improvements:

Barneskjerf “Softy” - jeansblå -> Barneskjerf “Softy” | Kløfri | Fri frakt & Klarna

The idea is to include some of the main properties about the item that “trigger” people to click and the additional information about free shipping and Klarna can also be a trigger.

Separating by | instead of - is also pretty standard to save on horizontal space and avoiding text cut-off.

Reviews

Currently no reviews on the products. This can be good for both SEO and conversion, but of course will take some time to build up, definitely advantages and disadvantages with setting it up. But longterm, done right, it is a net positive.

TBH judge.me is very cheap app and as good as all the other expensive ones, so I recommend it.

file naming convention

Stop using _ in filenames, use - instead. To quote Google:

“Use hyphens, not underscores, to separate words—for example, query-data.html. Search engines interpret hyphens in file and directory names as spaces between words. Underscores are generally not recognized, meaning that their presence can negatively affect SEO.”

Cashmere_scarf_Flow_cold_creme_3 -> Cashmere-scarf-Flow-cold-creme-3

Search console

Search volume is seasonal as expected, Kashmir is more relevant in the colder months in Norway.

Thoughts on SEO improvements to do

GA4 setup

Setup has been done through the Google sales channel and I cannot see that any mistake have been made.

If problems with session mismatch persists I would try to simply make a new account and connect again through https://admin.shopify.com/store/kashmina-no/apps/google

The Google sales channel should “work out of the box”.

Performance

The theme by default starts resolving a few domains, related to for example shopify product reviews etc. which are not used. Removed. This has a tiny, tiny impact, bust still adds some unnecessary latency if resource is never used.

unnecessary scripts and apps

Geolizr: When moving over to one store (shopify payments), can setup built in redirect based on country in Shopify markets settings

GDPR app: Snake oil, I would have just deleted it. Annoys customers and makes page load slower.

But, if you want a cookie banner, use the built in one

Then it will load faster as it will not be loaded from a third-party server. And also save 328 usd per year.

Some other apps like wishlist plus is “bad” for performance, but increases conversion, so is worth the trade-off IMO.

Theme settings

In theme settings things like “Show language selector” is enabled, but not used. Unchecked so less “work” is being done for a feature that is not used. If adding more languages it is of course fine to have these enabled.

SVG optimizations

In general SVGs are pretty small anyways, but running:

➜ Downloads svgo kashmiana-logo_1c9b81eb-ad36-4e6c-8324-372d862296a9.svg

kashmiana-logo_1c9b81eb-ad36-4e6c-8324-372d862296a9.svg: Done in 16 ms! 2.095 KiB - 16.3% = 1.753 KiB

So 0.3 KiB saved, not much, but still something.

Image optimizations

The lighthouse report about “next gen” image formats is a bit “outdated”, Shopify is already doing optimizations
serving images in better formats that the browser supports + optimized.

Also with regards to webp, due to improvements like jpegli

jpeg is actually “fine” for most images. Optimize.photos

is one of the few pages that use jpegli, so recommend using that one for image optimization.

Front page

The performance score on the front page is is not green like famme’s for example, mostly due to the high LCP, which is due to the video on the front page. If one can switch to a more lightweight image, it would probably improve the score somewhat, but if this decreases conversion, it might, since connection in Norway(main market) is generally good, I think it might be unnecessary.

Ideally one should load a mobile image for mobile and a desktop image for desktop, but that is a bit more work. For Famme we have made a custom image banner which let us upload a separate image for mobile and desktop. Then customer’s on mobile get a smaller image, which is faster to load.

Accessibility: Fixed some issues with missing aria-labels for screen readers. But in general there are some apps and theme elements that are not accessible.

SEO issues:

  1. a link from the upcart app which is not crawlable, easiest is to ask the app to fix this
  2. the “Mer” link in the header, ref. Google’s guidelines, google does not like links with “more” and similar as text, it should be more descriptive. Either expand out the menu or write something else like “Annet”

collection page

Performance:

404 for black.jpg and white.jpg swatches are missing in files section, has to be fixed, run through optimize.photos before uploading.

Lighthouse also complains about LCP here, one idea is to simply remove the collection banner image + description at the bottom of the collection and simply display the product list. This will probably improve the LCP score. Another thing if one wants to keep the banner, is to NOT lazy load since it is above the fold (visible on page load).

Loom explaining where to change lazy load on these images

I advise going through all the collections to ensure lazy loading is not enabled on above the fold banner images.

Product page

The SEO and accessibility is not getting full score here due to the same reasons as for the other pages, some links in the header and app links which are not crawlable.

For the performance, on most product pages it is decent, but apps + LCP (main image) of product page is main issue. I would recommend testing if these images can be a bit more compressed without affecting quality.

But around 89 which I get when testing most product pages is “good enough” not anything to worry about.

account pages

«account» pages til kundene (https://www.kashmina.no/account) så «ordre historikk/order history» blitt borte igjen.

added the section main-account.liquid back and used it in accounts.json template to show the order history.