Legal and necessary insider trading for marketing professional

Originally published in the June 2008 edition of teh New Zealand Marketing Association's DLB magazine - authored by Carlson Marketing

If your marketing budget was equivalent to an investment budget, trigger marketing is an investment strategy that seeks out high yield returns and invests only in those stocks (customers) rather than in every stock in the market.  By contrast, standard marketing invests only in the stock index (all

 potential customers in the market ) and hence each stock (customer) gets a piece of the investment budget.  Over time the index (all customers) does deliver returns which are both humble and directly proportional to the sum invested.  The only way to make more money investing in the index is to increase your investment (more share of voice at the customer). 

There is one critical distinction between stock investment strategy and marketing investment strategy – marketing professionals are legally permitted to be insider traders.  In fact – if they’re not insider trading then they’re not performing their fiduciary duty.  Insider trading delivers the holy grail of the investment professional – high return with low risk.  Trigger marketing delivers the same for marketing professionals.

Profile your customers. 

Trigger marketing, when used properly, is highly discriminating and mercenary.  We generally find four major customer tiers.  Tier 1 are our outstanding customers who are clearly giving us all their business in our category.  These will be less than 10% of your base and in general the goal with them is retention.  

Tier 2 probably comprises another 20% of the base and they’re giving us considerable business but are clearly shopping around.  We want to

 reduce their promiscuity.  These two tiers together are giving us close to 80% of our revenue.  The last two tiers have some value but the returns they generate aren’t as good as we can get out of the first two.

Each customer has a profile and a specific ‘shopping basket’ or purchase history.  This history builds a picture of who they are, and what people just like them (their peers) buy.  By comparing them to their peers we can identify 

what they’re not buying from us and by default must likely be buying from somewhere else.  We use this data to promote the products purchased by their like minded peers that they’re currently not purchasing from us. 

The net effect?  You know who is likely to respond to what kind of offers, and more importantly, what discounts you need NOT apply; reducing mass markdowns and allocating your marketing dollars to those who will respond best and deliver the most in return.

The key areas

There’s generally five areas for data gathering:

1 - Profitability

Look at the profitability of both customers and products and the intersection of the most profitable of each with the other.  You’ll also need to know individual product margins both list and realised after all discounts and incentives. 

2 - Basket typology

Identify the popular categories and their profitability.  Also understand what products are mixed in which baskets. 

3 – Promotional promiscuity

This is more difficult to assess as we find most organisations don’t retain data on customer responses to offers they’ve made to them.  However it’s one of the more important measures as it’s the measure of the impact of your past marketing efforts.  Most importantly you want to know who’s a discount only shopper.

4 – Life stage

Age and gender feature prominently here as expected.  Also useful is self reported information such as hobbies and interests but these are less valuable than actual behaviour as measured through facts.

5 - Shopping habits

Key here are the recency and frequency features most marketers will be familiar with.  However insights behind the data are also useful as tested through various hypotheses. 

Once the data is assembled the real work of developing a predictive model begins.  The five core data areas defined above are assembled into working ecosystem with each area feeding the others.  Done successfully a good predictive model will tell you with a high degree of accuracy (greater than 75%) who will visit which of your stores or branches next, when they’ll do this and what they will buy when they get there.   Much of our work in this space is now based on medical and actuarial survival models and the results for sections of the customer base are excitingly accurate.

For example, for one of the tiers in a retail client’s customer base we’re able to predict with an 80% level of accuracy when a customer is due back in their store. 

One of the benefits of trigger marketing is how well it fits with the other data driven areas of the business such as:

>        Finance - you can predict the revenue lift for a given campaign and demonstrate how much better than the next best option it might be

>        IT  - you can predict how many hits the website will take and hence the bandwidth required

>        Inventory management – you can link customer demand by category to your supply chain

The results speak for themselves.  The following are examples of successes we’ve had using this approach:

 

Client	Communication	Result
Credit card issuer	70,000 customers	Spend lift 43%, transactions lifted 67%
Upscale Department Store	497,000 customers	88% lift in  volume
Home Retailer	105,000 customers	48% lift in volume
Office Supply Retailer	185,000 customers	108% lift in volume
Hotelier	594,000 customer	$1Million incremental volume
Grocer	40,000 customers	132% lift in volume
Office Supply Retailer	185,000 customers	98% lift in volume

The principles of trigger based marketing are relevant to any business and the biggest returns come from the first rough and ready steps in treating different customers differently.