Ecommerce SEO consultant, with our own stores as the test bench
Updated
TL;DR: We run three online stores of our own, and we do SEO on them before we sell it to anyone. One store went from 79 Google clicks in its first three weeks to 1,281 clicks in its third month. It also made just two sales, both from paid ads, because almost all of that search traffic landed on pages nowhere near a purchase. That lesson is the core of how we work: we pick keywords by how close they are to a sale, we get new pages crawled fast, and we judge SEO by what happens on your product pages, not by a total-clicks chart.
What we do
Keyword and content plan built around buying intent
Before we write anything, every topic gets a simple label: someone ready to buy, someone comparing options, or someone just looking for information. Information pages can still be worth writing. But we decide up front where their readers go next, whether that's a product page, a related offer or your email list.
Technical SEO, within whatever your platform allows
Titles and descriptions, URL handles, sitemaps, structured data, redirects, internal links and page speed. On hosted platforms we work inside the limits the platform sets and tell you plainly which parts can't be changed (see the platform table below).
Getting new pages indexed
We submit new and updated pages through Google Search Console instead of waiting for Google to find them, and we decide which pages go first, because the daily allowance is small.
Reporting that ends at the product page
Monthly Search Console review broken down by page type, with clicks to product and checkout pages reported on their own line. If you have order data, we line it up against search traffic too.
GEO (AI search)
Structuring pages so they can be quoted by AI search tools like ChatGPT and Perplexity, with comparison tables and direct answers near the top.
Proof from our own stores
All figures are from our own stores, pulled from Google Search Console and our order database on 2026-10-07. The stores sell in Taiwan, in Chinese. The mechanics carry over to other markets. The market data doesn't.
Maomuse (custom pet portraits): traffic grew, sales didn't
| Period | Clicks | Impressions | Avg. position |
|---|---|---|---|
| Jul 12–31 (first 20 days) | 79 | 2,718 | 16.0 |
| August | 296 | 9,327 | 9.9 |
| September | 1,281 | 55,439 | 8.6 |
The site went from plan sign-off to launch in 3 days, with 62 articles and 12 tools live the day after launch. Our SEO audit score went from 54 to about 82 over successive rounds of fixes, and a check on September 27 found none of its 95 pages returning an error.
And still: two paid orders in total, NT$663 (about US$21), both from Meta ads, none after July 29. Over the last four weeks of data, the pet-name generator got 329 clicks. The page where you actually order a portrait got 28.
What went wrong is now our first rule. The top pages served people naming a new pet, researching end-of-life costs, or trying to comfort a friend. Those are real needs, but none of them is "I want to buy a portrait right now." We've since rebuilt the store's revenue model around the traffic it actually gets (digital files plus affiliate offers on the information pages). The full story, including what hasn't worked yet, is in the Maomuse case study.
Manual indexing: the one lever that worked immediately
New pages on Maomuse weren't getting crawled on their own. We submitted 9 URLs through Search Console's URL Inspection tool, and all 9 were recrawled within 13 minutes. A control group we didn't submit didn't move at all. One of the submitted pages went from position 94–98 to 30.6. Google's own documentation notes there's a quota for individual URL submissions; in practice we hit roughly 7–8 a day, so we rank which pages go first.
Taolilife (dried peaches): the main keyword never showed up
Our dried-peach store grew from 995 weekly impressions and 21 clicks (Aug 10–16) to 1,936 impressions and 48 clicks (Sep 19–25). But its main product keyword, "dried peach" in Chinese, got zero impressions from August 1 to September 27. The clicks came from a guide explaining the white crystals that form on dried fruit, and from local gift pages. On the speed side, we cut its mobile LCP from 13.1 seconds to 2.2: Chinese web fonts made up 76% of the homepage's transfer size, and switching to system fonts fixed it.
Platforms we support
Our own stores are custom builds. For hosted platforms, the table below comes from each platform's official documentation as of 2026-10-07. It isn't a record of our own hands-on work on these platforms, and "not documented" means we couldn't find it, not that it doesn't exist. We check inside your admin before we scope anything.
| Platform | What SEO can change | What it locks |
|---|---|---|
| Custom build (ours use Next.js) | Everything | Someone has to own hosting and maintenance |
| WordPress + WooCommerce | Everything, through SEO plugins | Same: hosting, updates and security need an owner |
| Shopify | robots.txt (via robots.txt.liquid, which Shopify calls unsupported customization), themes, sitemaps, redirects, URL handles | URL prefixes are fixed (/products/, /collections/) |
| CYBERBIZ (Taiwan) | Editable theme.liquid, so structured data, canonicals and head code can be added | robots.txt, URL slugs and hreflang aren't documented |
| EasyStore | Theme code is editable; URL handles for products, collections, pages and blog posts | robots.txt is controlled by the system; custom code has to be redone if you switch themes |
| SHOPLINE (Taiwan) | Titles, descriptions and descriptive URLs per page; built-in product schema; hreflang; redirects; an editable llms.txt | robots.txt is view-only; no code access, head code only through GTM; URL prefixes fixed |
| WACA (Taiwan) | Sitemap, SEO settings, tracking integrations | No blog on the entry plan, 5 posts on the next one; content SEO needs the 100-post plan or above |
| Wix | robots.txt editor, built-in structured data | Taiwan payment and logistics support not documented |
| 91APP (Taiwan) | Off-site work only (content strategy, reporting) | No public documentation; the platform sells its own SEO service |
On every platform, content, internal links and manual indexing are fully in your hands. On the more locked-down ones, that's where most of the gains come from anyway.
What we won't promise
- No guaranteed rankings or traffic numbers. Nobody controls Google's results, and our own stores show that traffic doesn't automatically mean revenue.
- No guaranteed sales from SEO. Maomuse is the proof: 1,281 clicks in a month and no new orders.
- No changes your platform doesn't allow. If robots.txt is locked, it stays locked; we'll tell you what that costs and work around it.
- No made-up first-hand claims in your content. If nobody has tested a product, we don't write as if someone has.
FAQ
- What does an ecommerce SEO consultant actually do for my store?
For us it's four things: choose keywords by how close they are to a purchase, fix what your platform lets us fix technically, get new pages indexed quickly, and report results on your product and checkout pages, not just total clicks.
- My traffic is growing but sales aren't. Is that normal?
It happened to us. Our pet-portrait store's clicks grew from 79 to 1,281 a month while orders stayed at two. Split your clicks by page type. If most land on information pages, the problem is intent distance, not ranking.
- How fast can SEO work on a new store?
Our pet-portrait store's average position went from 16.0 to 8.6 in under three months. A main product keyword on our other store still had zero impressions after eight weeks. Speed depends on how competitive your keywords are; we'd rather pick terms you can actually reach.
- Does manual indexing in Search Console really help?
On our store, 9 submitted URLs were all recrawled within 13 minutes, while unsubmitted pages didn't move. There's a daily quota, roughly 7–8 for us, so it needs prioritizing.
- Can you do SEO on Shopify?
Yes. Shopify gives more control than most hosted platforms: robots.txt, themes and redirects can all be changed. URL prefixes are fixed, and Shopify warns that robots.txt edits are unsupported, so we handle that file carefully.
- My store is on a Taiwanese platform like SHOPLINE or CYBERBIZ. Can you work with it?
Yes, within each platform's limits. SHOPLINE doesn't allow code changes, so the work is mostly fields, content and internal links. CYBERBIZ lets us edit
theme.liquid, so we can add structured data ourselves.- Do you only work with stores in Taiwan?
No. Our stores and platform research are Taiwan-based, but search intent, indexing and technical SEO work the same way elsewhere.
- How is pricing set?
Each engagement is quoted on your store's situation. Tell us what you sell, your platform and where you're stuck.
Want to know where your search traffic actually lands?
Send us your store URL and platform on Telegram. We'll start by splitting your clicks into "close to a sale" and "far from one."
Related: All ecommerce services · PPC consultant · Analytics consultant · Our three stores