It would seem obvious: if you send an email to a visitor of your online store, then bother to ensure it contains products that matching interests of each recipient, those he was looking at the website (or, at least, related products). Mass emailing without considering the visitor’s behavior and shopping path — is the loss of part of the profit. Obvious? Yes! Do all online retailers follow it? No!

I am subscribed to mailings of about 15 online stores, which I occasionally visit. And from all these stores I get emails – once a week, three times a week or even more often. And all those emails are standard and designed as any other "mass"-mailings, but none of them consider my interest in particular products of those sites. I was looking at the coat, but didn’t purchase it – in the email I get offers for T-shirts, sweaters, shorts, but no coat! I was looking at baby clothes, and received an email with a discount offer on goods for schoolchildren. And the list goes on. None of these stores bother to offer me the products that would be relevant to my interests. The only desire I (and most of you, I suspect) have is to unsubscribe, or simply delete such emails without opening them.

Our experiment

We decided to check in practice whether an email containing products in which a recipient was interested gives to the retailer better results than an email with some random goods. For this experiment, we used our marketing tool «LeadHit» as it collects and analyzes all the data about each site visitor’s activity (which was crucial for this experiment).

Four groups of emails were taken for the experiment, distinguished by the way of selection of offers contained therein:

• Group 1: Products in the email were relevant, i.e. corresponding to the interests of each recipient (based on his actions on the site);
• Group 2: Products in the email were casual, chosen at random;
• Group 3: Products in the email were popular, that is, among the best-selling online;
• Group 4: Products in the email were not available, only calls to return and make a purchase were made.

Three random sites were used for conducting this experiment: sporting goods store, furniture store, and a clothing store. Emails were sent to 6,000 recipients, 1500 messages within each group mentioned above. Emails differed only by product groups, and the rest (like design, etc.) was identical. Next, we calculated the percentage of emails conversion (that is, the number of actual purchases resulting from such emails, attributable to the number of delivered messages).

And here are the actual results we received:

1) Conversion for the store of sporting goods

Group 1: 17% conversion

Group 2: 7% conversion

Group 3: 8% conversion

Group 4: 3% conversion

As you can see, emails with relevant products gave the highest rate of conversion, while emails containing no products - is the lowest.

2) Conversion for the clothing store

Group 1: 21% conversion

Group 3: 11% conversion

Group 2: 10% conversion

Group 4: 4% conversion

The situation is the same: the email containing no products appear to be literally non-convertible, and the conversion of emails with the products relevant to interests of each site visitor has even exceeded the conversion of the same group of emails sent by the sporting goods store, resulting in 21%.

3) Conversion for the online furniture store

Group 1: 13% conversion

Group 2: 5% conversion

Group 3: 5% conversion

Group 4: 2.5% conversion

This site also performed well with emails containing relevant products, while the emails of the other groups showed the same low conversion results.

The figures above show that the conversion of an email with relevant products tailored to fit interests of each particular recipient is always about twice higher than the conversion of emails containing irrelevant or no products. Per numerous studies and market statistics, the average conversion rate of mass mailings is equal to appr. 1% or even less. And our experiment demonstrated - it could have been much more if the letter contained relevant products.

This experiment cost us $1,000,000

This experiment cost us money - from sending emails of each group (except for the group with the relevant products), a large percentage of the conversion has been lost. And more than 1,000 orders have been lost as a whole. In terms of the money we lost that way, the total “price” of our experiment was roughly $1,000,000.

To double-check the figures you can always conduct this experiment on your own website if you have sufficient time and budget. But we are sure that the results you get will be only slightly different.

Experiment-based conclusions

Based on the above, here is a few practical conclusions on how to structure email communication with customers:

• It is crucial that each email you send contains the products the recipient was interested in while browsing your site. Additional products may also be added, but only as supplemental.
• Do not forget to send an additional motivation (even your general discount, or a small gift) for the recipient to make a purchase.
• Per a study conducted by the company Marketing Sherpa, mailings with products which from time to time require replenishment gives 53%(!!!) conversion rate. This is a very high rate. Thus, periodically remind regular customers of the need to replenish stocks of products.
• The figures here speak for themselves: cross-sell recommendations or related/supplemental products added in the mailings increase the number of purchases by 20% compared with mailings having no such offers. A client bought the microwave? Offer him a set of fire-resistant ware. Bought a cell phone? He may need a case. It is more convenient to buy in one store, especially if it shows to the customer that it appreciates the customer’s loaylty.
• You should always propose additional items. For example, a customer purchased an iPhone in your online store. Maybe later he will be interested in the new version of the phone. So, you already have two reasons for sending a relevant email: first email offering the accessories, and, after some time, the second email offering new items in store.
• If you collect and process data relating to each visitor of your site, and make parallels and analysis by figuring out patterns for certain groups of visitors, you can make excellent predictions of what a purchaser may want in addition to what he is looking for here and now. And if you propose such items to this purchaser – he will most likely buy it. So, automate such predictions, include such items in your mailings, and they will definitely help you earn more.