We know what we want but we don’t know the size of the target group or opportunity.
We should have a definition of our desired target market from the problem recognition phase we can now build that criteria and determine the available volume.
Things that you will need to consider when determining selection criteria include:
Base or Inclusion criteria are the things that positively identify a customer, they are likely to include demographic information as well as other attributes like product holdings or good credit scores. Some examples include:
- Marital Status
- Homeowner, etc.
- Product Holdings
- Positive Credit Scores
Exclusions will vary by industry and organisation but some of the basics will include the following:
- Do Not Call (Internal Customer Flag)
- Telephone Preference Service (TPS – Government Register)
- Do Not Mail (also other channel specific – email, internet, banner ads, etc)
- Adverse Credit Risk Scores
- Product Holdings to Exclude (don’t offer if they already have it)
- Overseas Customers
- Goneaway (“Returned to sender”)
- Incorrect addressing details
- Exclude if offered ‘y’ widget in the last ‘x’ days
- Exclude from this campaign after ‘x’ attempts over ‘x’ days
See Customer Contact Framework for a much more detailed approach to contact rules.
Sometimes opportunity sizing can be far more involved and we may need to understand and predict the likelihood of something happening in the future.
Propensity modelling is an analytical technique that involves a systematic investigation of data to determine the likelihood of a future behaviour.
Propensity models are a very useful tool in the campaign management space and can help with every type of campaign – retention, cross-sell, up-sell and acquisition.
Management may be concerned with the level of customer churn (aka. turnover or attrition) and want us to plug the gap. An analyst would need to start by profiling the customers that have already left the organisation. Then they need to work backwards through the data to see if they can pinpoint any causal links that ultimately lead to the attrition. They do this by creating a scoring system for a number of past observations or events.
Some of the things that an analyst might look for include:
- Declining utilisation of a product or service
- Increased number of complaints
- Negative social media postings
- Branch closures or other local disruptions
- Payment or transfer request to another service provider
It is important to remember that you are only as good as your data. A big mistake that analysts make when building models is to ignore the timeliness of the data it will be applied to. You can’t predict a future event if you include the current month’s data in your model, because many organisations don’t have current data in their marketing databases. If a model takes a month to refresh and your data is a month behind then you need to build that timeframe into your equation and look for causal links at least 2 months before the event you are trying to predict.
Demographic and Psychographic Information
- Demographic – age, gender, income, education, occupation, location, etc
- Psychographic – attitudes, habits, interests, lifestyle, etc
The former is normally information contained in your customer database and can be complemented with segmentation information from organisations such as Experian (see Mosaic).
The latter can be found through market research, focus groups or even analysis of social media use. Both are useful for understanding and getting insights about your customers.
The ultimate result of this phase is to have a ballpark figure of the number of customers that fit the requested criteria. This information is then put into a business case to determine if following up this group is viable or whether the idea needs to be refined or shelved.
Don’t get in the habit of giving stakeholders a high-level figure for a target market. Build as much of the criteria as possible and advise that contact rules, etc may reduce that volume further when it is run in a live environment. All selections are dynamic and will change over time. You can read more on this topic by searching for “Customer Contact Framework”.