Customer Contact Framework
The Campaign Development Lifecycle helped you understand the best approach to manage a single campaign, but how does it fit in with the entire marketing programme? How do you set-up your processes so that communications are cherry-picked to be the most suitable for a customer and limited to the channels that they want? In most large organisations there are hundreds of marketing campaigns run every year and made available through multiple channels. Campaigns can be driven by people from different departments within the organisation – legal/compliance, product, marketing, brand, market research… the desire to communicate is never-ending.
What is a Customer Contact Framework?
A customer contact framework is the governance around how often we want to contact our customers and for what reasons.
To develop your own framework your need to plan which rules should be applicable under which situations.
I recommend 3 layers of rules:
- Mandatory Communications – Should not be limited in any way (Legal Terms & Conditions notifications).
- Trigger Campaigns – Low volume, timely campaigns that are based on an event. These need to have a less rigid framework.
- Regular Campaigns – Regularly scheduled, mid to high volume campaigns offering products and services from across the organisation.
The third level is the area that we want to focus on. There are several ways to do this and they are dependent on the size of the marketing programme. A key goal of the framework is the ability to change the offers made to a customer over time. Even with the worlds’ best predictive models we still aren’t going to know exactly what every customer wants at any given point in time. A customer contact framework should automatically try the next best offer if the first offer hasn’t worked for ‘x’ occurrences.
Technique 1. Build the rules directly into the Campaigns
All of the clever rules are referenced directly in the campaign selection.
This technique is best implemented when using campaign management software, so that the update of a single rule can propagate to all campaigns that reference the rule.
Rule Type 1. Product or Service Grouping
The campaign will reference rules for product and service groups. Great for varying messages over time and reducing over-solicitation.
Select anyone selected for a campaign involving a specific product or product group (e.g. an overdraft, or all credit products) in the last 3 months.
Rule Type 2. Campaign Specific Contact Rule
The campaign will reference a rule that selects anyone already contacted for this specific campaign recently. There is nothing sillier than making the same product offerings again and again to the same customer if they don’t respond. There is a point where you will lose more money trying than the returns gained from trying.
Select anyone who has been selected for Campaign ABC123 in the last 3 months
(This criteria also needs to select customers that were held-out as part of a control group).
Rule Type 3. Global Contact Rule
Select anyone who has been selected for any discretionary campaign in the last 30 days
(As opposed to a Legal or Obligatory campaign)
Rule Type 4. Campaign Specific Total Contacts Rule (Campaign Frequency)
This rule type needs to be based on an aggregate query which counts the number of times a single customer has been made the same offer, over the life of their relationship with you.
Select anyone selected for Campaign ABC123 more than 3 times
(This can exclude those selected for control groups as we are only interested in actual contacts).
I consider the last rule to be one of the most important if you want to maintain high response rates and consequently a profitable programme. I believe that every campaign should include this rule. I am often astounded how many times a single customer can be made the same offer across the lifetime of their relationship with an organisation. Some customers are included in the same campaign dozens of times over their life!!
Want to check how many times your customers have been contacted for each campaign?
It’s simple, just create an aggregate query against your promotion history table – something like the following:
Select SubQuery.Campaign_Code, Subquery.Number_Contacts, count(*) as Total_Customers
from (select a.Campaign_Code , a.Customer_ID, count(*) as Number_Contacts
from CAMPAIGN_HISTORY_TBL a where Cell = ‘Mailed’
group by a.Campaign_Code, a.CUSTOMER_ID) as SubQuery
group by SubQuery.Campaign_Code, SubQuery.Number_Contacts
order by SubQuery.Number_Contacts desc
The query results will display campaign_code, The number of times a customer has been contacted, and the number of customers that have been contacted at that frequency. You may need to tweak the code to suit your campaign history fields and I also advise that you first check the history table for duplicates by date, campaign and customer (in my experience you will often find duplicate entries).
Why all the rules?
My premise behind the use of these rules is that you will tend to see a lower response rate for each successive iteration of the same offer. At some point there won’t be enough of a return on investment to justify the spend. If you look at the chart below it is showing an example of the conversation rate for a banner ad in the financial services sector. The more the banner ad is displayed the less the response. From experience I believe that most campaigns follow this trend irrelevant of channel, and therefore it is important to vary the messages customers receive, as well as the frequency of overall contact.
The goal behind this rules approach is to start by promoting the most profitable offer with the highest propensity to respond, and if that offer is unsuccessful we need to swap to another offer over time. No predictive model is going to truly tell what every customer wants at any given time. Similarly, many customers will have a high probability of responding to multiple offers at a given time, but that doesn’t mean you should make all those offers.
The only person who knows what the customer wants is the customer.
What does a customer contact framework look like?
In this type of framework:
- Each column represents a different target group or segment.
- Each box defines run-order for the month (priority).
- The contact rules within each box allow the offers to change for each customer over time, as well as limiting the total number of contacts.
- All of the code is maintained within the individual campaign. By using a core “segment” in a campaign management tool it is easier to update the rules and have them automatically propagate to all campaigns referencing the code.
There always needs to be a balance between propensity and profitability.
Nearly everyone will have a high likelihood of wanting a free iphone if they open a Pay-As-You-Go phone account with you… but will you make money from it?
Technique 2. Specialist Campaign Optimisation Software
There is CRM software available that makes the ranking and sampling of campaigns a more streamlined, transparent process. It offers a variety of options that the user can tweak to manage a large number of marketing campaigns at any given time. They often make use of propensity models to determine which offers are the best fit for a customer. This software does make it easier to manage a large number of campaigns than that proposed in Technique 1 above but at a much greater expense. Generally you will need a dedicated resource to manage the software and process as well. It also means that you need to implement a more rigid framework when all campaign selections need to be run and ready by a set time. For absolute surety you may want to implement a combination of Techniques 1 and 2 above. Technique 1 (building rules in campaigns) will ensure that your contact framework is being followed even if it isn’t so obvious in the optimise software – a “belts and braces” approach (doubling the checks in case something goes amiss). For more information on this type of software consider reviewing the SAS Marketing Optimization software website here (opens in a new window).
If you want to understand how Optimisation software works you could consider trying to understand the Rank() function mentioned Here. This function can be used in SQL to order a list of customers by a propensity score and only capture those customers above a particular cut-off score.
Technique 3. Specialist Rules-Engine Software (Next Best Action)
There are a number of real-time rules engine software solutions as well. The general premise behind this software is that it will pick the right offer for each customer based on their individual characteristics and behaviour as they interact across various touch-points. For Example, if a customer were logged in to online banking and deposited a large sum of money this engine could immediately put forward a banner ad offering a competitive rate on a high-interest savings account. An option that traditional database marketers could only have dreamed of (data latency problems, etc). This software is great, but it still comes down to the cleverness of the algorithms that have been written. The timeliness of events is the key to the power of this software, but it doesn’t mean that all of a sudden you will be getting 100% conversion rates, for example most people with a large deposit already have the funds earmarked for something else. If you are contemplating a rules-engine approach you will need to devote a lot of time to the planning phase. Which rules should take precedence? Why? How can we tweak these rules over time? How do we manage them and make sure they can be phased out correctly over time? For a nice video to understand the concept more you could view the Pegasystems video here (opens in a new window).
Review invaluable tools for helping with the Campaign Management Function – Click Here