I launched MarketOnline in 2005, when many Romanians were still wondering whether it was safe to buy anything online. In 2009, in the middle of the financial crisis, I started Floria (an online flower delivery shop). Then, for years, I built Retargeting, a marketing automation platform for online stores, and saw from the inside how many shops work: some growing nicely, others burning money month after month without understanding why.
Friends and early-stage founders often ask me: “If you started an online store from scratch today, what would you do?” I decided to write the answer down properly, once. It is not a guaranteed recipe. It is the plan I would follow myself, in the order I would follow it.
1. I would pick a niche I understand, not an everything store
The Romanian market already has a few big players that sell everything. You cannot beat them on price, range or delivery speed on day one. But you can be the best in one corner of the market: gear for a specific hobby, products for a specific problem, an audience you know from the inside.
The question I would ask is not “what sells well?” but “where can I be the best?” If you know the product better than the customer does, you can advise them, write about it and answer the phone with real expertise. That is hard to copy. Floria worked partly because flowers are an emotional product with clear buying moments, which I learned very well.
2. I would validate demand before buying stock
The most expensive mistake I see is a warehouse full of goods nobody wants. Before the first big order, I would run a few cheap tests:
- look at search volumes for those products in Google’s free tools, to see whether people actually search for what I want to sell;
- run a few small campaigns with modest budgets to a simple product page, to see what a click costs and how many people reach the cart;
- try pre-orders or a waiting list: if people will not even leave an email address, that tells you something;
- start with small test batches or, where possible, dropshipping, so I do not lock cash into stock before I know what sells.
These tests will not tell you everything, but they tell you enough not to bet it all on a hunch.
3. The platform: simple at first, not perfect
Today there are good options for every budget. Hosted SaaS platforms such as Shopify give you a working store in a few days, with no servers to worry about. WooCommerce, on top of WordPress, gives you more freedom but also more technical responsibility. There are also Romanian platforms built for the local market, with ready-made integrations with local couriers and payment processors.
I do not think there is one right choice for everyone. I would choose whatever lets me launch quickly, connect couriers, payments and marketing tools easily, and export my data if I want to move later. The perfect platform does not exist. The platform you launch this month beats the ideal one you launch next year.
4. I would do the maths before spending on ads
Many founders switch on the ads and only then ask whether they make money on each order. I would do it the other way round. Before spending a single euro on advertising, I would know the answers to a few simple questions:
- How much is left from a product’s price after cost of goods, VAT and fees?
- What does delivery cost me, and how much of it does the customer pay?
- How many returns do I expect, and what does one return cost?
- What is the average order value, and how much can I afford to pay to acquire a new customer?
If a new customer costs you more than the margin on their first order, that is not necessarily a problem, but only if you know for sure they will come back. If you do not know, you have a problem. Say you sell a low-margin product with expensive shipping: maybe the answer is a bundle, a free-shipping threshold or simply a different product.
5. Delivery, returns and payment are part of the product
Customers do not just buy the item. They buy the promise that it arrives on time, well packed, and that they can send it back without a fight. I would treat delivery and returns as part of the product, not as a logistics chore. A clear, generous returns policy sells more than any banner.
On payments, cash on delivery is still dominant in Romania. Many people prefer to pay the courier at the door because that is how they trust a new shop. Card payments grow every year, and I would offer both from day one. But I would keep a close eye on cash-on-delivery parcels refused at the door, because they can quietly eat your margin.
6. First channels: SEO, Shopping, social and email from the first visitor
I would start SEO and content on day one, not “when we have time”. Product descriptions written by hand, not copied from the supplier, and articles that answer customers’ real questions. I have written in more detail about content marketing that actually sells. It is a slow channel, but it is the only one that gets cheaper over time.
In parallel, I would start Google Shopping, because it catches people exactly when they search for the product, and Facebook and Instagram campaigns for discovery, especially if the product is visual. And I would collect email addresses from the very first visitor, with their consent. Email marketing is not dead, and a list built early is an asset nobody can take away from you.
7. Retention and automation, not just acquisition
This is where most money is left on the table. Most visitors leave without buying, and many who buy once never come back, simply because nobody reached out to them again. That frustration is why I built Retargeting: the idea that a store has to bring its people back with the right message at the right moment.
In the first month I would set up a few basic automations: messages for the abandoned shopping cart, a thank-you email after the order, a review request after delivery and a complementary product suggestion a few weeks later. You do not need a big budget for this; marketing automation is no longer a luxury, even for small shops.
8. Brand and reviews
In a niche, people buy from whoever earns their trust. A name that is easy to remember, a clear voice, real product photos and a phone answered by someone who knows what they are talking about. Reviews are the best advertising: I would actively ask for them after every delivery and show both the glowing ones and the less flattering ones. A shop with nothing but five-star reviews looks suspicious.
9. What I would not do
- I would not spend a big budget on design before the first sale. A clean, fast, clear site is enough at the start. Design gets refined with money from sales.
- I would not launch with too many products. Better 50 well-described, well-positioned products than 5,000 copied from a feed.
- I would not launch without tracking. Analytics, conversions, the source of every order: if you do not measure, you are just guessing. And guessing is very expensive.
- I would not copy the big players’ prices. In a niche you win on expertise and service, not on the lowest price.
If I had to sum it all up in one sentence: start small, somewhere you can be the best, do the maths before the ads, and look after the customers you already have. The rest comes with patience.
What would you do differently if you started an online store today? Or, if you already run one, what do you wish you had known on day one?