The first ninety days of selling online as a manufacturer

Why the first ninety days are mostly not about the website

Ask a manufacturing team what is needed to sell online and you will usually get a list of website features. Ask the same team for the shipping weight of every SKU in its despatch carton and you will get silence, a spreadsheet from 2019 and an argument about whether the figure includes the pallet. That gap is the project. A competent ecommerce build takes a predictable number of weeks. Assembling accurate, complete, consistently structured product information for a few hundred SKUs takes longer than anyone expects, and nothing can launch without it.

The second thing that sets the schedule is operational. Someone has to pick single units in a facility organised around bulk despatch. Someone has to print labels, book couriers, answer a customer asking whether the bracket fits a 2018 model, process a refund and decide what happens to a returned item. In most factories those people already have full time jobs. If the plan does not name them, the launch date is fiction.

Product data is the actual project

Every downstream system reads product data: the store, the marketplace listings, Google Shopping, your freight rate calculator, your accounting system and your warehouse. If the source is inconsistent, every one of those systems is inconsistent in a different way, and you spend the next two years reconciling them by hand. Build one authoritative record per SKU, owned by one named person, and export from it. That single decision saves more money than any platform choice.

• Decide the single source of truth in week one: the ERP, the store, or a dedicated product information tool
• Validate before import. Missing weights and duplicate SKUs are cheap to fix in a spreadsheet and expensive to fix in production

Field	Why it matters	Common failure
SKU and internal part number	Ties the order back to production and stock	Two numbering systems that nobody reconciles
Product title, written for search	How customers find it	Internal model codes used as titles
Short and long description	Conversion and search relevance	Copied from a supplier PDF, duplicated across listings
Structured specifications	Filtering, comparison and buyer confidence	Specifications trapped inside a PDF datasheet
Variant attributes	Correct options at checkout	Colour and size baked into separate products
Despatch dimensions and weight	Accurate freight quoting	Retail box measured instead of shipping carton
Dangerous goods classification	Carrier eligibility and packaging	Discovered at the depot after launch
GTIN, MPN and brand	Marketplace and Google Shopping eligibility	Missing, so listings are suppressed
Compliance marks and certifications	Legal and buyer trust	Held by the quality team, never published
Stock location and lead time	Realistic despatch promises	Made to order items shown as in stock
Images and asset references	Every channel needs them	Files named IMG_4471 in a shared drive

Photography and copy you can reuse for a decade

Product photography is the one asset that outlives platforms. Sites get rebuilt, marketplaces change their rules, but a clean, consistent, well lit library of images will still be working in ten years if it is shot to a specification rather than to a mood. Shoot once, shoot properly, and shoot to a written shot list so that product number three hundred looks like product number one.

• A shot list per product: hero on white, three quarter angle, scale reference, key detail, in use, packaging, what is in the box
• Shoot at a resolution high enough for zoom and for print, then export web sizes from the master
• Capture dimension diagrams and exploded views for anything technical. Industrial buyers use them
• Video only where it answers a question, such as installation, assembly or a durability demonstration

Own store, marketplace, or both

The instinct is to launch everywhere at once so nothing is missed. Doing that in the first ninety days usually produces three half finished channels, inconsistent stock and a support inbox nobody owns. Pick one channel to be excellent at, prove the operational model, then add the second once the pick, pack and service process survives contact with real volume.

The honest framing is that a marketplace rents you demand and an owned store builds an asset. Marketplaces put you in front of buyers immediately and teach you what sells, at the cost of margin, control over the customer relationship, and rules that can change without notice. An owned store costs more to fill with traffic and pays back over years through repeat customers, email lists, organic visibility and data you keep.

• If you go the marketplace route, plan Amazon or eBay listings from the same product data source as the store, never as a separate spreadsheet
• Do not sell the identical SKU at a lower price on a marketplace than on your own store. It trains customers away from your highest margin channel

	Own store	Amazon or eBay	Both
Time to first order	Slower, demand must be built	Fast, demand already exists	Fast, if operations can cope
Margin per unit	Highest	Reduced by referral fees and fulfilment	Mixed
Customer relationship	You own it	The marketplace owns it	Split
Data you keep	All of it	Limited	Partial
Rule risk	Low	Listings and policies can change	Diversified
Best first move for	Differentiated products and spare parts	Commodity or accessory lines	Teams with existing despatch capacity

GST, ABN and the tax setup before your first order

Most established manufacturers already have an ABN and are registered for GST, so the tax question is usually about configuration rather than registration. If you are not yet registered, registration is required once turnover reaches the ATO threshold, and it is generally sensible to register before launch rather than mid quarter. GST in Australia is ten per cent, and the practical consequences show up in three places: how prices are displayed, what your invoices contain, and how sales land in your accounting file.

Display first. Consumers must be shown a single total price that includes GST, prominently, so a store aimed at the public should be configured to enter and display prices inclusive of tax. If you also sell to trade customers who expect prices excluding GST, the platform needs to switch display by customer group, and not every platform does that well. Check it during evaluation rather than after the build, because it is a common cause of a rebuild.

Then invoicing. A tax invoice has required content under ATO rules, including your identity and ABN, the date, a description of what was sold, the GST amount or a statement that the total includes GST, and for larger sales the buyer's identity or ABN. Your store's order confirmation email is not automatically a compliant tax invoice, so check what the platform actually produces. Finally, if you export, sales of goods exported from Australia can be GST-free where the conditions are met, which is a configuration question in the tax settings rather than something to handle manually.

• Check that order confirmations meet tax invoice requirements, or issue invoices from the accounting system instead
• Agree with your accountant whether orders sync as individual invoices or as a daily settlement summary before anything is connected

Freight zones and shipping rules that do not lose money

Australia is a hard freight market for manufactured goods: long distances, concentrated population, and vast low density areas that carriers surcharge. Rates are built around zones radiating from the origin depot, and they price on the greater of dead weight and cubic weight, where cubic weight is commonly length by width by height in metres multiplied by a conversion factor. A light, bulky product is therefore priced as if it were much heavier, which is why carton design matters commercially.

1. Export every SKU with despatch carton dimensions and weight, and flag anything oversized, fragile or classed as dangerous goods
2. Get quotes from Australia Post and at least two other carriers, and read the surcharge schedule as well as the rate card
3. Run your real order and postcode mix through each quote to compare total cost, not headline rate
4. Choose the rate method: live rates for varied catalogues, table rates for a few profiles, flat rates only for uniform small goods
5. Set a free shipping threshold from the model, with exclusions for oversized items and remote zones
6. Test the checkout with genuine remote postcodes across WA, NT, TAS and regional QLD before launch

The internal operations changes nobody budgets for

This is the section that decides whether month four is calm or chaotic. Selling online adds a small, continuous operational load that does not fit neatly into an existing manufacturing rhythm. Orders arrive at all hours including weekends. Each one needs picking, packing, a label, a despatch scan and a tracking notification. Customers ask questions before they buy and expect an answer the same day. Refunds have to be approved, processed and reconciled. None of it is difficult. All of it is relentless.

The systems side matters just as much. If the store and the inventory system do not share stock, you will oversell within a fortnight, and overselling a made to order item is a worse customer experience than showing it as unavailable. Decide how stock synchronises, how often, and what happens when the connection fails. If the volume justifies it, connect the store to the ERP through a documented interface rather than manual export, and use workflow automation for the repetitive parts such as tracking notifications, review requests, backorder alerts and low stock warnings.

• A named order owner and a named backup, with a published despatch cutoff time
• A physical pick and pack station stocked with cartons, void fill, tape and a label printer
• A support inbox with a one business day response standard and template answers for the top ten questions
• A returns bay, a returns workflow and a rule for what happens to returned stock

The ninety day plan, week by week

This sequence assumes a catalogue of a few hundred SKUs, an existing brand, and a team doing this alongside their normal work. It front loads the slow items and leaves the genuinely fast work until the end, which is the opposite of how most projects are scheduled and the reason most of them run late.

1. Weeks 1 to 2: pick the launch range, name the data owner, agree the source of truth, and write the field definitions. Start the SKU audit
2. Weeks 3 to 4: complete data for the launch SKUs including despatch dimensions and weights. Get carrier quotes. Choose the platform and get a fixed written quote for the build
3. Weeks 5 to 6: photography against a written shot list. Write product copy for the launch range. Begin the build and the analytics setup
4. Weeks 7 to 8: configure freight rules, tax settings and payment. Connect stock and accounting. Draft policies covering shipping, returns and warranty
5. Weeks 9 to 10: load products, check every page on a phone, and run the checkout end-to-end including remote postcodes and a refund
6. Weeks 11 to 12: soft launch to staff, dealers and a small customer list. Ship real orders. Fix what breaks
7. Week 13: open publicly, tell the channel, and start measuring contribution per order rather than revenue

What to leave until after launch

Scope creep is the main reason these projects slip from three months to nine, and most of the ideas raised in early workshops are easier to specify once real orders exist. Things worth deferring: the full catalogue tail, customer accounts with saved configurations, trade pricing tiers, a product configurator, marketplace expansion, subscription consumables, a loyalty programme and multilingual content. Things not worth deferring: accurate freight, correct tax settings, stock synchronisation, working analytics and a returns process. The first list can wait a quarter. The second list causes losses from day one.

After launch, the highest return work is usually unglamorous. Improve the product pages that already get traffic, fix the checkout steps where people abandon, and write the technical content that answers the questions arriving in the support inbox. That is where conversion work and eCommerce SEO pay for themselves, and it costs far less than another platform feature.