The cost of maintaining a CRM can be large, however the cost of getting the data wrong can be immense.
Data8 have estimated that the cost of preventing an error at source is about ten pence. The costs immediately start to escalate as the error is left in the system.
The error could have been made by the customer or a member of staff but either way anything that will ensure the accuracy of the data will save at best £5.00 and at worst £50.00.
The incorrect data can hang around a CRM database for years causing problems with poor communications and complaints leading additional admin costs to follow up and reply. If this is identified early it may cost as much £5.00 per record to correct.
To quote the philosopher George Santayana, “Those who cannot remember the past are condemned to repeat it”, so if the error is not corrected it could cost £50.00.
So how to achieve this –
If you are setting up a CRM database and importing data or already have a system that has been in use for some years it will pay to get the data cleansed. Reducing the number of errors will improve the efficiency of the database and a quick data cleanse can be done online and overnight. This will produce a free Data Quality Report, allowing you to see what the current state of your data is. It may also be the time to look closer at the data held and get a single customer view.
One of the biggest source of error is duplicate records, if you are using some de-duplication suites they will only match where there is an exact match. It is important to use a system that contains ‘fuzzy logic’ allowing you to capture more duplicates e.g. Bentley and Bently or Data8 Ltd and data 8. Using our bureau service you can have access to a number of algorithms that will allow you to define which data to merge automatically, which to discard and which to flag for manual intervention.
By removing goneaways and duplicates you will improve the quality of your data and start to see how many customers you really have.
Having achieved a cleansed data base the next stage is to maintain it. Data is not in a static state, new data will need to be input and rules need to be in place to ensure that data is not added any old way or you will quickly end up where you started.
As soon as the data is cleansed it will start to decay, therefore anything that can help the accuracy of any future data input will be of benefit. In some cases, for less than a 1p, you can validate, in real time, the postal address, email address and telephone numbers. Rules should be established within CRM which won’t allow any further details to be completed until the data is valid. This ensures consistency of data as the CRM database grows.
Also there must be compulsory data fields that must to be completed and completed in the correct format, this could be driven by a set of drop down boxes or regular monitoring of the data quality in certain fields. The more ‘free’ fields there are the greater the risk of getting incorrect data or not being able to get any measureable metrics as everyone will have their own idea of what to include.
The data is now clean and any further additions will meet with some strict guidelines. Now is the time to look at enhancing the data.
If you are business to business retailer or service provider, maybe you should consider adding not only telephone numbers and emails but a range of business information to get a fuller picture of your client. Data8’s Business Profiles service uses market leading data sources like Companies House, who update their data daily. We also provide quality, comprehensive, credit reports including company name and number, credit rating and limit, P&L details, balance sheet details and directors.
For consumer customers we can add telephone numbers and emails to addresses, the addresses themselves can be segmented into different types of demographic profiles. This can be used to target any messages correctly and help identify where similar types of individuals are concentrated.
To quote Eddie Cochran, following these three steps should send you to CRM data heaven