When customer knowledge precedes planning
Few marketers attempt to plan without the support of some form of understanding of customer behaviour; nevertheless a very large proportion plan with partial, or more dangerously, unsound, insight support.
This is frequently because the wrong questions have been asked of the support analyst, but also because the customer data has not been assembled to enable consistent insight results. Useful requests for insight stem from the need to make better planning decisions: e.g
- picking the right mix of digital marketing spend requires an understanding of how to attribute digital sales back to customer journeys
- undertaking a successful CRM upsell campaign needs knowledge about when, why, and which customers repeat purchase
Customer data is not always easy to ‘assemble’; it may be residing in multiple silos, updated at different time frequencies, and with no universal match-key to tie it all together. Very few organisations achieve a true single customer view.
This short paper outlines a journey you can take when empowering marketers with the insight and tools they need for successful marketing planning.
Our definition of ‘true customer insight’ is the complete set of tools required to inform the decisions you need to make when planning marketing.
This in itself is a tall order when every different business environment requires its own mix of insight; contrast what you need to know when using Google AdWords with organising door-drops!
You also don’t want to be overloaded with tools; we believe that the majority of insight work is not productive because it is not directly linked to planning decision making. Useful insight changes the way that actual marketing decisions are made.
And the insight needs to be derived from as complete as possible a customer view; this may need to cover all or some of transactions, registration information, digital journeys, inbound and outbound contacts, triggers, and response history.
Yet at the same time the customer view needs to be as simplified as possible to support the insight to be derived from it; big data can be a curse when for instance transaction summaries are all you need.
In one case we found that, having trawled through millions of credit card transactions, there were only a couple of transaction types that were useful as variables in a model to predict insurance risk
Here are, however, some examples of the direct links between insight tools and making planning decisions:
- understanding which channels to use, and how to target recruitment, depends on understanding the different levels of longer-term customer value they provide
- providing the right on-line content can depend on interpreting interests displayed in customer browsing journeys round the internet; for offline content, research results from a customer panel may be required
- putting together the best product mix in home shopping catalogues, for the right audience, requires a customer segmentation that forms customer groups from the similarity of their product purchases
- knowing whom best to approach with an offer on a customer base depends on both understanding via propensity models their level of affinity for that product, and with an understanding of whether there are other products for which they have a greater affinity
- getting the timing right for CRM activity requires a knowledge of when customers typically make a second purchase, and when they are likely to attrite.
1 Using a data audit to inform our clients as to how to strengthen the single customer view
The data audit starts with joining, where possible, all existing sources of customer data to scope the potential for a single customer view, and providing counts of the numbers of active and inactive customers. We will test matching on all possible keys including name-address, emails, telephone numbers, and account numbers.
This enables an understanding of the inter-relations between data sources, and pins any data quality issues back to their source, so that improvements can be made to data capture activities.
We next profile the customers using available internal data such as tenure (time since joining), and geographic distribution. As part of this, the audit looks at the potential for creating derived variables such as gender from title, and recency, frequency and value from transactions etc.
Apart from providing a description of the customer data asset, this process allows us to specify in detail how the single customer view can best be constructed.
2 Developing or enhancing the single customer view
From the outset of direct marketing, the common issue facing all marketers was deduplication. Originally, the main cause of this was addresses, due to their inconsistent formats, splitting and merging of house numbers and names, people moving, and differences in the end users’ writing styles.
Initially data bureaus developed match keys or similarity matching programs. Nowadays there are various tools available at relatively low cost rates, such as Cygnus or Matchit SQL enabling users to process their own files. In the case of Cygnus this also includes PAF validation and suppressions such as the National Change of Address file (NCOA).
As time has moved on channels have increased, but these channels have fixed formats such as telephone numbers, email addresses and social media alias. For most marketers the new fixed formats have reduced the need for deduplication. It is now often the case across many sectors that email addresses are used as account numbers/URN’s.
But how unique is this new unique reference number, and why is this important?
The ‘why is this important’ used to be simple; direct postal mailings were expensive, de-duping and suppressing a file at say 5p a record would save £0.50 a record; a clear cut cost-saving business-case. However emails are now cheap – so why are there still benefits from an SCV?
When looking directly at digital communications costs there is little benefit. However, the real wins are in the customer relationship value; deriving insight tools, personalised marketing, customer value development and contact density management are all useless if we cannot identify an individual.
We work with our clients to build or enhance their single customer view, and to fix where possible any flaws in the way that customer data is collected at source.
Frequently we end developing an analytical data mart (or ADM). The driver behind an ADM is to provide analysts working to build insight tools with a consistent, unified, stable and reliable single customer view. This will often be a single table with all the key customer variables included. The ADM can be used for instance for developing customer segmentations, customer value measures, and customer propensity models, and then for holding the resultant values at the customer level.
3 Building the insight tools required
As we have seen above, an essential step is to design the toolset required for the planning needs of the enterprise.
Most organisations have some of the tools, in varying states of repair, but few have all.
The very large companies that have their own analytics team can progress to fill the gaps; the rest require outside resource to help them fill them.
One often overlooked question is, how to measure the ROI derived from insight tools, and then how to prioritise them? The ROI comes from the extent to which marketing performance can be improved as a result of using the tools.
If better digital results attribution leads to a reduction of AdWords spend of £X, for the same level of sales, we have a very tangible result. Similarly if a set of propensity models can improve demand from an overall catalogue budget for a home shopping company by £Y, we also have a measurable result.
Of course before the tools have been built, and when they still need to be justified, we can only look at the scale of the area affected, and the likely level of improvement.
Developing certain insight tools requires a prior step of undertaking some marketing tests; this is particularly true in areas like understanding the impact of customer density on customer responsiveness. Where tests are involved, costs and opportunity costs are much greater, but we believe that an ROI can still usually be measured.
Once agreement is reached on the toolset to build, then implementation can be done in several ways. At Berry Thompson we often mentor and train people inside client organisations to build insight tools. Alternatively we can undertake the build ourselves and apply the resultant scores and values to the single customer view.
4 Time to start planning!
Encapsulated within the concept of ‘marketing planning’ are a series of key strategic decisions to be made around:
- brand development
- financial objectives e.g. sales
- product development and pricing
- customer recruitment and attrition levels
- deployment of distribution channels
- marketing budget requirements and justification
The marketing planner also has to deliver
- a strong ROI from marketing expenditure
- detailed allocation of the marketing budget to specific marketing activities
- effective use of marketing and sales resources
- support for corporate strategic objectives such as product launches
Our overall approach looks like this;
Most organisations we deal with plan at the middle and bottom layer of this model.
The annual planning cycle deals with investments in marketing and sales programmes, customer journey stages, and on-line and off-line channels.
The implementation planning deals with expenditure on actual marketing campaigns and the detail of how they are to be designed.
Berry Thompson can advise on the overall marketing planning process, based around the organisation’s objectives combined with insight derived from customer knowledge.
Alternatively we can help you plan individual on-line and offline campaigns, including financial planning, target setting, channel mix and choice, content, and results monitoring.
We do this using our holistic model of the marketing planning process combined with relevant insight tools.
To be effective this becomes a joint process where you set the objectives, and we plan the means. And to be sustainable the cycle repeats each period.
You can reach a virtuous circle when all links in the marketing planning cycle are joined together.
Every organisation we deal with has some of the links in place, but few have them all joined up and working.
Whether you want help with a single element, such as the data-audit, or development of insight tools, or a wider review of all planning processes, then please contact us.
Our resources comprise marketing planners, statistical insight developers, and data engineers experienced in building single customer views for marketing planning.