A/B price testing
Setting pricing strategy for the products sold is sometimes very difficult, because there is no ready set of rules, which would help to do it in an easy way. In this post we’re focusing on the situation, when the same products are available in different e-shops.
So what is the definition of ‘optimum price’ for a product? It’s a price thanks to which we’ll achieve the greatest profit from sales of a particular product in a definite period of time. It means that the aim of it is profit maximization, which is defined as the amount of products sold x margin on 1 piece in a definite period of time.
the amount of products sold× margin on 1 piece in a definite period of time = profit maximization
We’ll simplify it. We assume that the main factor, deciding about the product purchase in a particular shop is price. It’s only partly true. Depending on a country, the percentage of people, for whom the price is the most important criterion of a shop choice oscillates from 60- 88%. So we can assume that the price is the most important criterion of choice.
So how to define ‘optimum price’?
You can do it effectively, pure and simple, carrying out price testing and analysing an influence of price change on the sales (profit).
How to prepare test:
1. Choose a product, which has a big rotation from the perspective of a test period. I suggest that the minimum week rotation should be >10 pieces ( in the case of tests, which last about one month).
2. Set an initial price level – remember to set it in relation to competitive prices: we set price on the + 2% level from the cheapest offer in the market or an arithmetic mean from the prices of 3 cheapest competitors. We keep such a price for 2 weeks (of course if competitive prices change we also change our price).
3. We raise/lower the price about 2-5% for a period of one week.
4. We go back to the price level from point 2 and we keep it for one week.
5. We repeat the process (points 2-4) 3 times.
There are many factors that influence on sales level (except price). Our test is going to give chances for getting realiable data with error risk minimization.
During next iterations, we can raise the level of price change and search optimum price in this way.
1. A typical A/B testing carried out on one product may be badly perceived by clients. You can carry out tests on similar products or use ‘disposable special offers’.
2. Remember not to carry out tests during time of a great demand fluctuation, for example during clearance sale.
3. Products, which are chosen to first price experiments should have quite big rotation, preferably rather constant. There can of course appear some seasonal fluctuations (short and long-term).
4. Always compare prices in relation to competitive prices.
5. You can carry out lowering prices tests, using special offers.
6. If you have products that your competition doesn’t have, you can pattern your prices upon similar products.
7. If you have historical data, you can start from analysing it. Check if there is a relation between price change and changes of sales levels. On the basis of it, you can try to define level of change.
8. Calculate the exact margin – take into consideration all costs, for example: packing, shipment and so on. It’s especially important in the case of low-margin products. Correctly calculated margin is a base of calculating profit from sales.