Many businesses use leads to help grow their customer base. Purchasing marketing leads gives businesses an opportunity to expand to satisfy growth targets. Whilst many businesses opt to benefit from this opportunity, few do it effectively. In this blog, we list a few very important initiatives that would assist to ensure effective leads management.

 Effective use of data

A marketing list purchased from an organisation registered with the Direct Marketing Association (DMA) will provide you with a number of potential customers. Many organisations will try and contact each name on that marketing list, however, the quality of each lead is unknown and the response rate is likely to be very low. This is where data (with models) play an important part. Data used by organisations to help improve leads’ quality and response rates include the following:

1. Opt-out list – this is a list of customers who have asked not to be contacted. Your marketing list provider may apply this automatically or you may need to apply it yourself. There are many opt-out lists, including one from the DMA.
2. Existing customer data – is the customer an existing customer? You may or may not want to contact them. If you are willing to contact your existing customers, you need to determine whether they are in good order. Filtering your leads through your existing customer database is therefore a necessary first step.
3. Internal black-lists – businesses should keep lists of previous customers with whom they no longer wish to do business. This would include fraudulent accounts or written-off accounts, for example.
4. Credit bureau data – this is particularly pertinent for those offering credit or incidental credit products. Avoiding reckless lending is important and it is therefore critical to know whether a potential customer is under debt review, has a fraud flag against his/her name, has excessive judgements, excessive defaults or a very low credit score. Credit bureaux are rich in data and have other useful data such as geospatial data (perhaps you want to know if a customer is located near one of your stores) and contact data (that can be used alongside the marketing list data).

Data is crucial in developing intelligent leads that can help improve the response to your campaigns and the quality of business onboarded.

2. Utilising models

What makes leads “intelligent”? It’s the use of data and models. Using models in lead management can dramatically improve your response rates as well as the risk profile of your lead. Multiple models are used in the world of “intelligent prospecting”. The first model (which can be used regardless of the product you are selling) is the response model. This model measures the likelihood of a customer responding to an offer from you. You can thus save time by not focussing on customers with a very low probability of taking up your product, thus saving you money.

Along with response models come risk models. For credit and incidental credit products, these are models that predict the likelihood of a customer defaulting on the loan or facility that you offer the customer. For products such as insurance or fibre contracts for example, lapse/attrition models are also used. It is expensive to originate a new customer. If that customer lapses that policy after only a couple of months, originating the customer would most likely have cost the organisation money.

Populations do change over time and sophisticated organisations should look to improve their models regularly utilising machine learning approaches.

3. Waterfalls

The utilisation of waterfalls is a very important part of the leads processing process. Here a large base population is extracted. From this base population, individuals are excluded based on a variety of criteria. It might be that they are an existing customer, they have an unacceptable risk profile or maybe they are on the do-not-contact list. They are then excluded from the final candidate list. Once the waterfall is applied, the models are applied and the selection can take place.

Waterfalls are personalised for each lender, based on internal policies and requirements. The waterfall process is applied for each selection (e.g. monthly).

4. Execution – aligning origination strategies

After a waterfall has been applied and one is left with a final candidate list, response and risk models need to be applied. Once a list of customers with acceptable risk profiles has been created, the customers need to be contacted. These customers are often still subject to additional scorecards, policy rules and affordability assessments. Such risk assessment steps typically utilise data not known prior to engagement with the potential customer. Where data is known, it should be brought into the waterfall. The scorecard (acquisition and originations) should also be aligned to ensure that the risk profile is consistent. There will naturally be swap-sets, but it is important to ensure that this is as small as possible. A drop-off of 5-10% (after applying an origination scorecard) is reasonable.

For affordability, some organisations conduct extensive affordability calculations and also request proof of income (POI) data. A drop-off of 30% (sometimes as high as 50%) is not uncommon following affordability assessments.

5. POPIA related initiatives

Since the enactment of the Protection of Personal Information Act (POPIA), only customers who have opted in to digital communication (e.g. email and SMS) may be contacted through these channels – it is therefore important to keep this regulation in mind. We have noted that marketing teams now execute separate strategies for digital leads versus “analogue” (e.g. call-centre or mailed) leads. The opt-in example above and other POPIA requirements have resulted in marketing teams introducing a number of initiatives to ensure POPIA compliance.  

6. Channel optimisation

Engaging the customer through the correct channel is an often-overlooked step within the acquisition process. Utilising a digital strategy (where possible), an analogue strategy or a combination of the two might bring different results, depending on the consumer you are contacting. Channel models utilise information around age, gender, LSM group and credit/incidental credit product holdings, to better equip marketing teams in selecting the most appropriate channel to engage the potential customer.

Furthermore, not only is it important to choose the correct channel, but also call centres to get better responses when contacting potential customers at the correct time. This optimal time can differ from customer to customer. Leading marketing teams utilise right-time-to-call models to better optimise their call centres, thereby saving time, reducing cost and increasing contact rates.

7. A/B or Champion/Challenger Testing

Whether running call-centre campaigns or digital campaigns, best practice is to continuously test what you are doing. For call-centre campaigns, you may want to measure one leads’ provider against another or compare different risk and response models. You may want to test calling customers utilising right-time-to-call models versus calling a customer at random times. Testing allows us to better understand the effectiveness of what we are doing.

For digital campaigns, testing one’s messaging by sending out two (or more) different messages on randomised segments of the population will assist in understanding which message was most effective. Prudent marketing teams will extend this further by testing when the best time to send out a campaign would be. All campaigns need to be effectively tracked.

8. Accurate reporting

It may seem obvious, but many businesses do not have adequate acquisition reporting. Leads reporting should look at several metrics including:

• Campaign performance – How well did the campaign perform? What was the eventual take-up rate by channel? How many times was the lead contacted?
• Model performance – is the actual response rate in line with the expected response rate? Is the eventual risk of the accounts onboarded in line with the expected risk? Is the lapse rate in line with the expected lapse rates?
• A/B testing – as mentioned earlier, effective campaigning means constantly testing what you are doing. For this to occur, effective tracking is crucial. We would recommend running “difference reports” where key metrics are compared on each group to determine the most effective campaign.

9. Call centre agent KPIs

In order to improve campaign results, motivation of call-centre staff is essential. Setting effective key performance indicators (KPIs) and tracking these through the day encourages sales staff to work hard to convert the opportunities. Virtual agent assistants such as Principa’s Agent-X, enable the agents to keep up to speed with their sales performance. Further to this, gamification is also possible, financially incentivising agents on achieving certain targets.

10. Existing customer marketing

Whilst new customers are a major focus of marketing teams, it is also good to focus on the existing customer base. It is well known that retaining a customer is more cost effective (by at least 5 times) than trying to lure new customers to your business. Therefore, growing your existing base by cross-selling and up-selling should be an active part of your marketing strategy. The list of initiatives above talks to many things that are as pertinent to existing customer marketing strategies as they are to new customer marketing: strategies, data, models (for example spend models, anti-attrition models and propensity models), channel optimisation, champion/challenger testing and reporting all play important roles in trying to grow revenue associated with your existing customers. All of this should be automated and utilising a decision engine such as Principa’s DecisionSmart can help you realise this.

In the retail space, adding merchandise analytics where you market specific products to individuals based on their merchandise spend, is an advanced level of sophistication that top lenders are exploring. The critical components required are access to this data and the ability to execute the strategies. Principa’s Stratus machine learning platform is a solution that enables this.

Conclusion

Whether you have an advanced analytics team helping you conduct your marketing campaigns or are just using raw marketing lists to attract new customers, there are always incremental improvements that you can introduce into your business. Principa has been working with several organisations for over twenty years helping them improve their marketing strategies through many of the initiatives mentioned above. If you would like to learn more about these initiatives and the opportunities they present, please do get in touch.

Author – Thomas Maydon, Head of Credit and Analytics 

Thomas has 18 years experience in data science focusing in the EMEA region. His experience traverses multiple industries and disciplines covering analytics, consulting and software solutions for companies ranging from large banks and retailers to telcos and manufacturing operations. A large part of his experience comes from working with credit providers helping them harness the predictive power of data through the use of machine learning and decision tech. Thomas holds an MA in Mathematics from Edinburgh University.