The name of this retail segment is self-explanatory – “Fast Moving Consumer Goods” implies frequent and regular purchasing. A customer is bound to come back for more and often sticks to a stable shopping schedule.
It gives FMCG stores an outstanding opportunity to develop a sustainable sales flow. Reactivate your customers monthly, keep the sales flow at a stable level – sounds easy. Is it also that easy in reality?
Smart Assistant instead of a Generic Recommender
The primary goal of marketing personalization is to grow the conversion rate and AOV (Average Order Value). Any personalization is based on complex mathematical and statistical algorithms that process the gathered data about your visitors. A lot of solutions offer you a Big Data approach.
This situation makes the personalization in a lot of stores follow the two case scenarios.
The first case scenario is “Frequently Bought Together”:
A visitor views a product page or adds a product to the cart. The on-site product recommendations generate the purchase candidates, for example “Garlic, fresh champignons and 15% cooking cream are frequently bought with this pasta”. This is a logical choice meant to help the visitor cook a delicious pasta.
That kind of personalization requires manual sorting and linking when the store first forms groups of interconnected products: champignons and pasta in one group, tea and lemons in the second, the third group for chicken, spice mixes and parchment paper.
The second case scenario is “Customers Who Bought This Item Also Bought” and illustrates the Big Data approach:
If a group of customers bought this sort of shrimps with that sort of coffee (e.g. due to a big discount on both at that time), every new buyer of this sort of coffee will be recommended these shrimps. Doesn’t sound very logical? Well, the statistical data says it’s been done thousands of times, so why wouldn’t it work for every new buyer?
Big Data will keep confusing your customers with these awkward recommendations, but you’ll still see an increase in sales. After all, the statistics are correct (partially).
It is the right moment to shift your attention to the missing factors. The first significant factor FMCG stores miss is a purchase frequency. Consider you have a regular customer, who loves shopping for groceries, vegetables, and fruit, those taking more than 50% of each new purchase. This customer is shopping right now and has just added to the cart yummy Cluster Tomatoes. Your store immediately starts recommending various sorts of pasta, cheese, sauce and other ingredients from the recipes that include tomatoes.
But instead of stimulating regular purchases and recommending the other items this person usually buys, the store offers generic recommendations irrelevant to the customer:
What is hindering the recommendation system in this store from being a real smart assistant to the customer? The reason is obvious – the recommendation algorithms are unable to remember the individual purchase history and recommend the customer relevant products. Instead, the system makes general assumptions and gives no genuine, personalized customer experience.
In other retail segments, the situation may be different. For example in Consumer Electronics, it is not logical to recommend another laptop to a future laptop buyer. Or another laptop bag. In this product category, a time span between the purchases is usually very long, and a much more effective tactic is to personalize the recommendations to show a future buyer good quality accessories. Same as in Apparel & Accessories, where you would recommend a pair of cute red shoes to go with an elegant red purse, not some red winter boots.
However, FMCG is a different story. In Fast Moving Consumer Goods a purchase frequency is a key parameter. There, specific recommendation systems alone, having a niche solution for each retail niche, can provide an adequate level of service.
Here is a clear explanation of the underlying mechanics and logic of a correct type of personalization driven by the Progressive Personalization technology.
Purchase Frequency Analysis
The method to deduce purchase frequency is clear and simple: the system runs a purchase history analysis for each of your customers, for the last month. Then it filters out irrelevant data keeping the records for the items bought more than once by your customers.
| Product | Week #1 | Week #2 | Week #3 | Week #4 |
| Liquid soap | 1 | 1 | ||
| Wholegrain bread | 1 | 1 | 1 | |
| Toilet paper | 1 | 1 | ||
| Washing-up liquid | 1 | 1 | ||
| Alkaline batteries | 1 | 1 |
Based on these records, the system makes a hypothesis about the next month purchases. From this moment on, it recommends relevant products at an appropriate time to each of your regular customers.
Progressive Personalization and Effective Email Marketing Campaigns in FMCG
Employing the following techniques, you’ll be able to increase and sustain your sales flow.
Trigger Emails “Regular Purchase.” Unlike emails, on-site recommendations are not smart enough to remind your customers about upcoming shopping sessions. Sending emails with personalized content will cover both your active and lapsing audiences. Alternatively, use instant Web Push Notifications to maximize the readability (no clicks-to-open required, native to each internet browser).
You have a choice between sending single-product emails or group-product emails. Single-product emails usually work well for other retail niches, as FMCG has a rapid product flow and it’s easy to tire the customer with too many single-product emails.
On the other hand, the group-product emails are designed to help your customers with their regular shopping plan. Armed with the data, the system optimizes upcoming shopping sessions offering a checkout-ready shopping cart for a group of regularly purchased products. The customer can pay on the go and, thanks to you, will never run out of the supplies.
For instance, each 10-15 days a regularly shopping customer gets a reminder email with a checkout-ready shopping cart.
This table shows the items and quantity from a pre-formed shopping cart offered to the customer shortly before the shopping session.
| Products | Week #5 | Week #6 | Week #7 | Week #8 |
| Liquid soap | 2 | |||
| Wholegrain bread | 3 | |||
| Toilet paper | 2 | |||
| Washing-up liquid | 2 | |||
| Alkaline batteries | 2 |
An important detail this table does not fully illustrate is that the number of days is not a static variable but is dynamically determined in real time, based on the purchase history of the customer.
You can always change the look of your emails to conform with the current brand image and colors, via a visual editor made by REES46.
Progressive Personalization + Big Data
Big Data-based personalization is a powerful, however imprecise tool. It doesn’t include into its computations important details about the visitor and suffers from the cold start problem. In other words, it wastes your money for a long period of time requiring weeks to gather and process the initial data.
All these flaws can be made up to when Big Data is coupled with Progressive Personalization – a tech proved by recent years performance, applied by many successful mid-sized and big online retailers.
Progressive Personalization is in demand because:
- It doesn’t require tons of traffic from the store. It fits well, even the stores with the number of unique visitors below 30,000 daily.
- It starts giving an adequate performance from the first on-site clicks of the visitor.
- It constantly gathers the data and progressively adapts the user experience, in real time.
- It relies on the key parameters of each visitor (inсl. gender, age, income level, relationship status, number of kids in the family, brand preferences and much more).
You can couple Progressive Personalization with Big Data or use it as an independent solution. In both cases, you will shortly see a substantial increase in revenue, conversion, AOV and your other KPIs.
You are free to use the described logic and mechanics in your solution or apply a final product, for a quick and easy integration.