What has to be ready on launch day for a new online store to take orders in its first month instead of its sixth: catalog data, product pages, checkout, stock sync, Merchant Center, analytics and ads. Based on a building-materials store I built solo, from design to the ad campaign.
- Published
- Reading
- 14 min
- Author
- Yaroslav Tsarenko

Most new online stores follow the same script. The design gets signed off, products get uploaded, the site goes live, and then nothing happens. The owner waits for search to "kick in", turns on ads, sees clicks but no orders and has no idea where people drop off, because analytics was going to be set up "later".
This guide is for owners of small and mid-sized businesses who are about to launch a store, or who have one that does not sell. I'll walk through what has to be ready on launch day for orders to start in the first month rather than the sixth. The examples come from a recent project: a building-materials store I built on my own, from the first design screen to the ad campaign.
Why a new store stays quiet for the first month
There are usually three reasons, and none of them is the design.
- Nobody can see the store. Organic search takes months to warm up for a new domain. Without a Google Merchant Center feed and search ads, for the first month the store exists only for people who already know the URL.
- The store doesn't answer the buyer's question. A shopper sees the price of one bag but can't tell how many bags they need or what the total will be with delivery. They leave for a site where that is obvious.
- Nobody knows what's happening. Without ecommerce events in analytics, ad campaigns optimize blind and the owner can't see which step loses people.
How much is lost at the last step is well documented. Baymard Institute's average across 50 studies puts the cart abandonment rate at 70.22% (Baymard). Leaving out people who were "just browsing", the top reasons US shoppers give are extra costs that are too high (40%), slow delivery (20%), not trusting the site with their card (19%) and being forced to create an account (18%).
Almost none of these are fixed by advertising. They are fixed by decisions made during the build: show the full cost early, let people check out as a guest, be honest about delivery times. So a store that sells in month one doesn't start with a campaign. It starts with everything below the waterline.
Below the waterline: what a store that sells is made of
Shoppers see the storefront: home page, catalog, product page, cart. What actually sells is everything underneath. Here is how that looked in my latest project.
Parandis Bud is a building-materials store in Khmelnytskyi, Ukraine, and an official Ceresit dealer. I built it end to end by myself: the design, the store on Next.js 16, the admin panel, shipping through Ukraine's two national carriers (Nova Poshta and Ukrposhta), fiscal receipts, stock sync with BAS (accounting software from the 1C family, widely used in Ukraine), the Merchant Center feed, analytics and the Google Ads search campaign. How I covered every role myself, and what went wrong along the way, is in the story One Developer Instead of an Agency.
None of these numbers are decoration. Each one is a separate decision. Without any of them the store would look exactly the same and sell worse. Let's go through them one by one.
Shopify or a custom store?
The first question owners ask is why anyone would build a store from scratch when Shopify, WooCommerce and dozens of hosted platforms exist. The honest answer: often you shouldn't.
Hosted platform
- Live in daystheme, payments and shipping out of the box
- Simple catalogproducts that need no calculations
- Inventory in the platformintegrations limited to its app store
Custom build
- Your business logiccalculators, wholesale, complex pricing and shipping
- Stock from your accounting systemthe ERP is the source of truth
- Full controlspeed, structured data, feed and analytics
A hosted platform is a great fit when the catalog is small, products are simple (apparel, accessories, gifts), payments and shipping are standard, and you're happy to manage inventory inside the platform. You launch in days and pay a monthly subscription plus apps.
A custom build earns its cost when your business logic doesn't fit a theme. For building materials there were several such places:
- Material calculators. The buyer doesn't need the price of one bag. They need the number of bags for their job.
- Stock lives in the accounting system. Quantities and prices are maintained in BAS, and the site had to read them from there, not from a spreadsheet someone updates by hand.
- Very different order sizes. A construction crew and a homeowner redoing a bathroom buy differently, yet both have to be able to order on the same site.
- Control over structured data, the feed and speed. These decide how the store looks on Google before anyone clicks.
If none of these apply to you, start on a hosted platform and spend the money on photos, product copy and ads. If two or more apply, the template will start fighting you within the first year.
The catalog: data matters more than design
Parandis Bud has 449 products in 66 categories. The longest part of catalog work isn't layout. It's data: titles, attributes, units, photos, brands and codes.
Google Merchant Center always requires an id, title, description, link, image link, availability and price. Brand is required for all new products, and Google strongly recommends a GTIN (the manufacturer's barcode number) whenever one exists, as long as you're sure it's correct (Google Merchant Center).
What that means for an owner:
- A product title is a search query. "Ceresit CM 11 tile adhesive, 25 kg" gets found. "CM11 25" doesn't.
- Units are consistent across the whole catalog. If one product is priced per kilogram and the next per bag, neither the calculator, nor the filters, nor the feed will work correctly.
- Categories follow how people search. Not "Group A dry mixes", but "Tile adhesive", "Plaster", "Filler".
- Photos and specs come from the manufacturer or from you, never at random. A product without a photo or weight neither sells nor passes feed review.
One tip that saves weeks: start building the product spreadsheet before design starts. The data decides which filters, attributes and page templates the store needs.
A product page that answers "how much do I need?"
People buying building materials arrive with a specific job: tile 12 square meters, level a wall, prime a floor. They need an answer to "how much should I buy and what will it cost in total".
That's why six material calculators run on Parandis Bud's product pages. The shopper enters the parameters of their job and the page works out how much material they need. That removes two of the reasons from the chart above at once: not seeing the total cost and the fear of buying the wrong amount.
The principle isn't limited to building supplies. Every niche has its own "how much do I need":
| Niche | Buyer's question | What the product page should show |
|---|---|---|
| Building materials | how many bags for my area | a consumption calculator and the total |
| Wallpaper, flooring, tiles | how many rolls or boxes | a calculation from room dimensions |
| Pet food, cosmetics, consumables | how long will it last | usage period and a subscription option |
| Clothing and shoes | which size fits me | a size chart with measurements |
| Electronics and parts | will it work with what I have | compatibility and specs |
Then there's trust. Official dealer status, the warehouse address, a phone number, the returns policy and real delivery times belong on the product page, not just in the footer. In Baymard's data, 19% of shoppers abandoned because they didn't trust the site with their card, and a new store with no reviews yet is especially exposed.
Checkout, shipping and stock: where orders get lost
Cart and checkout
The cart and checkout are where it's easiest to lose someone who has already decided to buy. A few rules I build into every store:
- Guest checkout. Offer an account after the purchase. Mandatory account creation drove away 18% of shoppers in Baymard's data.
- Shipping is chosen on the site. On Parandis Bud both national carriers are integrated into the store, so the buyer doesn't have to wait for a sales manager's call to find out how they'll get the goods.
- The full total before the last step. The shipping cost, or at least its range, has to be visible before the shopper presses "Place order".
- Receipts and documents are part of the system, not manual work. In Ukraine, fiscal receipts are a legal concern, and on Parandis Bud they're integrated into the store. Wherever you sell, settle with your accountant how receipts and invoices must be issued before development starts. It decides which integrations to plan for.
Stock and admin: never sell what you don't have
The most expensive mistake a new store can make is taking an order for something that isn't on the shelf. The customer waited, a manager apologizes, the click that brought them is already paid for, and Merchant Center may flag the product if availability in the feed doesn't match the site.
That's why Parandis Bud syncs stock with BAS: the accounting system is the source of truth, and the site, the feed and the admin panel read from it. The admin panel handles what accounting doesn't know about: descriptions, photos, categories, calculators and order statuses.
Four things to check in any admin panel before you sign off:
- Can a manager change a price, photo or description without a developer?
- Where does availability come from, and how quickly does it update on the site and in the feed?
- Are all orders visible with contact details, shipping method and payment status?
- Is there a change log showing who changed a price and when?
Google Merchant Center and structured data
Merchant Center is where Google gets your product data for both ads and free listings. Free listings can appear on the Shopping tab, in Search, Images, Lens and YouTube in the countries Google supports, and Ukraine is on that list (Google Merchant Center). For a new store this is one of very few ways to get product views without an ad budget.
The Parandis Bud feed contains 461 offers. The one rule for any feed: build it from the same data as the site, or sooner or later the price and availability on Google will drift away from the storefront.
The other half is structured data on the pages themselves. Google states plainly that providing both structured data and a Merchant Center feed maximizes eligibility for rich results, and that some experiences combine data from both sources (Google Search Central). Parandis Bud uses 15 structured-data types, and each one describes something that is actually on the page.
The rule is the same one I described for service businesses in a website that gets leads when AI answers: markup must describe what a person can see. Made-up reviews or ratings in the markup lead to manual actions, not stars in the results.
Analytics before the first ad dollar: GA4, GTM and Consent Mode v2
Ads without analytics means paying for clicks without knowing what they bought you. So analytics has to be live before the campaign starts, not after.
Google Analytics 4 defines a set of recommended ecommerce events: viewing a product list, viewing a product, adding to cart, beginning checkout, adding shipping and payment info, and purchasing (Google Analytics). The purchase event requires a transaction ID, a currency and a value. Without them, revenue in your reports will be empty or double-counted.
Parandis Bud tracks six GA4 ecommerce events. That's enough to see the funnel from product view to purchase and to feed real conversions into Google Ads instead of "thank-you page visits".
Now, consent. In November 2023 Google updated Consent Mode to version 2, adding two parameters, ad_user_data and ad_personalization, to the existing ad_storage and analytics_storage (Google Tag Platform). Since March 2024, both new signals must be granted for EEA users' data to be used for personalized advertising (Google Display & Video 360).
If you sell into the EU, this is not optional. If you don't, there are still two practical reasons to build a consent banner and Consent Mode into the store from the start:
- Retrofitting tracking after launch costs more than building tags in Google Tag Manager around consent from day one.
- Advanced mode keeps part of the data. In basic mode, if a visitor declines, Google receives nothing at all. In advanced mode, tags send cookieless pings, and Google models conversions with an advertiser-specific model rather than a general one (Google Ads).
Google Ads in month one: search first, not everything at once
Once the catalog, the feed and analytics are ready, you can turn on ads. For the first month I recommend starting with a search campaign. Someone typing "buy Ceresit tile adhesive Khmelnytskyi" already knows what they want; the ad only shows them where to buy it.
The Parandis Bud campaign has 178 keywords in 13 ad groups. The point of ad groups is simple: each one collects closely related queries so the ad matches the search and lands the person on a page where they immediately see the product they asked about, not on the home page.
What did the first month bring? Per the client, a customer who found the store through Google Ads placed a single wholesale order of UAH 250,000 with the manager in the first month of ads. One order isn't a statistic, and I won't present it as a typical result. But it shows exactly why everything above matters: ads can bring a buyer to the site, and the buyer only orders if the catalog, prices, availability and checkout already work. Once wholesale orders like that become regular, they need their own flow, and I've covered what a B2B ordering portal needs in version one separately.
Rules for the first month:
- Negative keywords from day one. "DIY", "reviews" and "instructions" bring clicks without purchases.
- Read the search terms report every week. It shows what people actually typed before they clicked.
- Optimize for purchases, not clicks, as soon as the first conversions accumulate.
- Not everything at once. Shopping campaigns and Performance Max make sense once the feed is clean and you can see which products sell.
What a store that sells in month one costs
I delivered the full Parandis Bud cycle, covering design, the Next.js 16 store, the admin panel, shipping, fiscal receipts, BAS sync, the Merchant Center feed, analytics and the search campaign, for a fixed $5,000. One person was responsible for all of it, so there were no hand-offs between contractors and no "that's not our part" arguments.
Here is what the budget of any project like this is made of, and what makes it grow:
| Part | What drives the cost |
|---|---|
| Design | the number of unique page templates, not the number of products |
| Catalog | who prepares the data (you or the developer), how many attributes and filters |
| Product pages | calculators, variants, wholesale pricing |
| Checkout | shipping and payment methods, receipts, documents |
| Integrations | accounting or ERP system, carriers, CRM |
| Visibility on Google | feed, structured data, technical SEO |
| Analytics and ads | GA4 events, GTM, Consent Mode, campaign structure |
Budget separately for ad spend, the money you pay Google for clicks, and for your team's time on photos and product copy. How to compare quotes from different developers and what to look for in an estimate is covered in detail in MVP development cost in 2026; the logic for a store is the same.
Launch checklist
| When | What must be ready | How to check |
|---|---|---|
| Before design | a product spreadsheet with titles, units, brands, photos | every product has a price, a weight or size and a photo |
| Before design | decisions on shipping, payments and receipts | your accountant has approved how receipts are issued |
| Before launch | guest checkout and the full total before confirmation | a test order on a phone in under 2 minutes |
| Before launch | availability from the accounting system | change stock in accounting, see it change on the site and in the feed |
| Before launch | a Merchant Center feed without errors | products passed review |
| Before launch | product markup that matches the visible text | Google's Rich Results Test shows no errors |
| Before ads | GA4 events, including purchase with transaction_id | a test purchase shows up in reports with its value |
| Before ads | a consent banner and Consent Mode v2 | debug mode shows the consent state |
| Month one | a search campaign by category, negative keywords | a weekly review of search terms |
| Month one | a report of orders by source and product | you know which clicks made money |
The technical layer, meaning speed, server rendering and caching catalog pages, is invisible to the owner, yet it decides whether a product page loads on a phone over a weak mobile connection. A study by Deloitte commissioned by Google found that a 0.1-second improvement in mobile site speed went together with an 8.4% rise in retail conversions and a 9.2% rise in average order value (Deloitte). How I handle that in Next.js is in my App Router production guide.
What to do this week
If you're still planning a store, start with the product spreadsheet and a conversation with your accountant about receipts. If your store is live but doesn't sell, check three things: whether the buyer sees the full total before confirming, whether the purchase event fires in GA4, and whether your products are in Merchant Center. Most of the time the problem is one of these.
I build online stores and business websites from design to ads, and you talk directly to the person who writes the code. More about how I work is on the website development page. If you'd like to talk about your store, write to me: I'll look at what you have now and tell you what I'd fix first.
Was this useful?
– likes · – views
Questions people ask
How much does it cost to build an online store?
It depends on the catalog, the integrations and who builds it. As a reference point: a building-materials store with 449 products, custom design, an admin panel, two carriers, fiscal receipts, stock sync with the client's accounting system, a Merchant Center feed, analytics and a Google Ads search campaign cost a fixed $5,000 for the whole cycle. Budget for ad spend separately.
Should I use Shopify or build a custom online store?
A hosted platform is the right call when the catalog is small, products are simple and you are happy to run inventory inside the platform. Custom development pays off when you need product calculators, complex pricing or shipping rules, sync with an accounting or ERP system, and full control over speed, structured data and analytics.
Can I show my products on Google for free?
Yes, in the countries Google supports for free listings. Products from Google Merchant Center can appear on the Shopping tab, in Search, Images, Lens and YouTube at no cost. You need a feed with the required fields: id, title, description, link, image link, availability and price.
Do I need Consent Mode v2 for my store?
If you have visitors from the European Economic Area and use Google Ads or Analytics, yes: since March 2024 Google requires consent signals, including ad_user_data and ad_personalization, to use EEA users' data for personalized advertising. Outside the EEA it is not formally required, but it is far cheaper to build in at launch than to retrofit.
How long until a new online store gets its first sales?
Organic search alone usually takes months for a new domain. With a Google Ads search campaign and a clean Merchant Center feed, first orders can come in the first month, provided analytics is live before the ads, product pages answer the buyer's questions and checkout does not lose people.