Non-transparent model of ad settlement
I lost my first serious money on a marketing campaign based on the CPC + CPL model (cost per click + cost per lead). To my horror, it turned out that instead of the CPL (cost per lead), it was the CPC that generated most costs (cost per click). The advertising network was displaying our ads on a mainstream social medium, which resulted in lots of clicks and zero purchases.
No customer base
It took us a whole year to realize that it is easier to sell something to people, who already know about us. We wasted this year by not investing in the creation of the base of potential customers. Newsletter and social media turned out to be the most effective sales channels, with ROI much higher than in the case of other media.
Too many products in the offer
Long Tail strategy seems to be an interesting idea, while the Drop Shipping model allows for unlimited extension of the offer. We assumed that such an approach would help us to achieve better exposition in Google results and that it will make the up-selling easier. Unfortunately, the intense use of Drop Shipping is usually associated with the risk of losing control over the logistics. This way, our customers needed to wait a long time to receive their products. When I was running my e-store, the DS model allowed us to introduce large amounts of products in our offer. However, even our suppliers did not have some of them in their storehouses. This brought about incredible problems with the order management.
Not enough interaction with the customers
At the beginning, we used to subscribe to various internet forums and mailing lists related to house and garden decorations. We did that to ask a direct question – what do you lack? It turned out to be a great idea and allowed us to find some niches, for instance lamps in Japanese style. After introducing the product in our offer, all we had to do was to reply to our question, thank for the advice and inform people that the products they wanted are available at our store. When we developed our offer, we did not invest enough in the Category Management and we eventually lost the knowledge about the needs of our customers. The suppliers will always try to convince us that their product is a hit, but it is your task to feel what people need.
We used osCommerce to create the first version of our store. We modified this software ourselves and despite of low quality of the code, everything was quite stable. Unfortunately, the following version was created with the use of a dedicated solution. We thought that such software would be 100% compatible with our needs and therefore it would constitute the most convenient solution. However, this type of software is usually less stable and more problematic. Instead of sellers, we became testers and lead programmers. The time we needed for re-testing and bug reporting discouraged us to develop our system. We also had problems with the programming resources availability. In most cases, dedicated software makes it impossible to recruit independent developers. Good Open Source software would be a much better choice.
Beginning of the cooperation with the success-fee model
Each e-commerce manager dreams about finding an e-commerce agency that would boost the sales in exchange for a commission fee only. Obviously, companies of this type exist in the market, but many of them work for commission fees due to the lack of alternatives. Usually, negotiations of such contracts take much time and result in the performance of ostensible actions that do not have much impact on the business results. Only after several such cases I finally understood that good companies select contracts that are safe and that they do not want to bear the responsibility for the entire business of their client. With an appropriate scale of business, it is always possible to configure the cooperation in the success-fee model, but it is usually a better idea to begin with standard models and to invest in the advanced negotiations only after some time, given to both companies to test each other.
Time horizon too short
My first major AdWords campaigns were terminated, because I could not see their results after couple of days. I thought that the money I spend should return to me in the same month. Only after some time I realized that it takes time for the statistics to work and for the customers to make a purchase (sometimes after several visits). This scheme was recurrent. Investing in product descriptions, videos, content marketing or customer service always repays after some time. The amount of this time depends on the character of the given enterprise and constitutes the main business risk.
Today, the e-commerce market is much more competitive, therefore feel free to learn from my mistakes :).