How to link data to a CRM

It is the inevitable stakeholder question which follows every single segmentation study. The ability to replicate the results of a survey style data onto CRMs is data enrichment at its finest. It is nothing short of marketing heaven. Being able to link softer measures, such as attitudes and values to purchase behaviour and contact information is nothing short of a gold mine. It is the best question a marketer can ask about data strategy.

Many see survey data about their customers and imagine a world where they can port this information directly onto their CRM, to guide content and product suggestions. They know they could skyrocket the business, by merely tweaking their messaging and propositions - to speak to each customer as an individual, to target their underlying needs. Given half a chance, they will even blow their research budget on a consultant who offers them a few months of this make believe - a glimmer of hope that this might be a possibility. One must imagine that this delusion is worth it because it happens so frequently - there is a thriving industry around it. It is a heaven worth hoping for.

Data enrichment is a murky topic. It arrives with the undertones of the Cambridge Analytica scandal. It is illegal to enrich a customer database using information not gained from your customers through positive opt in. It is also not good enough to let a third party do the dirty work, because GDPR does not allow the transfer of responsibility. This heaven need not be about data enrichment, nor does it require creepiness.

Don't be creepy with Data Integration

Customers simply do not want you to know more about them than they have shared with you. They don't want you to meet them, then go find out more about them, without their consent, like a stalker. Its unnerving - not unlike walking into a car show room and the salesman already knows exactly who you are, and what you are thinking. Except in most cases, the sales pitch is coming straight to your inbox or through your letterbox. It's perceived as being manipulative.

The reason it doesn't work without data enrichment is because a little bit of purchasing information, and your customer's postcode doesn't tell you anything about them. The concept is based on the hope that if I know who you are, what you bought, or where you live - I know why you bought it. I appreciate how easy and perfect it sounds. It's a real idea! But it's also from the 1980s when attitudes, demographics, needs, and values stuck together. Not anymore. Not in a digital world. Most would now regard the very idea as offensive! It's certainly not what most marketers would consider to be on-brand or even in keeping with their original intention; to appreciate every customer as an individual.

The hard truth is that despite the obvious commercial rewards, the dream of an enriched CRM is beyond reach for most organizations. Not because of GDPR - but for the lack of vision, consistency, and commitment. For those who have failed to execute any data strategy around their CRM, the information they hold on their customers is merely a repository for a few fields of contact information. Sitting on such a CRM is tough and it makes the idea of instant data enrichment just that more salivating.

Strategic Data Integration: What to do instead?

The idea of being able to understand customers at a very deep level remains laudable and worthwhile. Though the objective remains the same, the path to get there is somewhat different, taking years of planning, committed data strategy underpinned be a genuine drive to build trust with customers - for the purposes of delighting them and doing good by them, not manipulating them into buying more. Some have hit this sweet spot. What unites them?

• They want to get to know their customers to ensure a product-market fit - for their customers. The goal isn't to enrich their CRM. It's a by-product. Data = Money. Your customers know that just as well as you do. It has to be a give-take relationship.
• When asking about soft touch needs and perceptions, they are extremely focused on their product and service. They are not greedy with what they ask and don't ask creepy things for the sake of completeness; questions that they might have no business knowing about, like "Are any young hildren living with you?" Tell me, what business is that of yours exactly? How does this change your targeting? What business is it of yours how much I earn? Do you suspect I can't afford your product?
• They communicate why they want the information - and are clear about how it benefits their customers.
• Their data collection is not haphazard, rather it is hardwired into their customer life cycle/journey.

You know that some companies get this right - you're probably one of their customers.

The limited nature of these sorts of techniques doesn't quite satisfy most marketers - who are looking for a more rounded understanding of the customers they have in their CRM. The upshot of this is that they build more relevant information into their insights process by design, opening the door to stronger predictive algorithms. This is where Data Fusion comes to the fore, as it is not limited by the amount of information which it can bring into the CRM. It just needs a proper foothold.

The goal here is not to attempt to target individuals. No predictive algorithm should, at the time of writing at least, give you enough confidence at an individual level that you could hope to creep out your customers through targeted messaging. The goal is to just be a little better at knowing who your CRM customers are, and to know who they are more likely to be. Surveys are still the best way to do this - because they don't rely on trying to infer anything through digital analytics, (although I'm sure this would have worked in the 1980s if the internet were around then).

Here is the winning recipe:

Step 1: Review your CRM. What do you know about your customers? Is it creepy, manipulative or none of your business? If so, move on to the next thing.

Step 2: Narrow the list to only the sort of information that would be helpful to build product market fit, reduce churn or improve customer satisfaction. Agree on the list. Get someone with clout to sign it off. It's permanent.

Step 3: Conduct a very small survey amongst your customers at a certain time in their journey, e.g., after remaining a subscriber for 3 months, incentivize the survey with rewards (which have nothing to do with your product and service, vouchers from premium national retailers work best). You must heavily incentivise participation this, as only your most loyal, delighted customers are going to give up their time for free - you are far more interested in knowing about the rest - those who will only be lured in with a pot of gold. Keep your questions focused on your business outcomes. Keep your survey down to 20 general data points only - don't be greedy, this is not a panel survey. Response rate is all important. Ask for a positive opt-in to merge their survey information with your CRM. Explain exactly what and why you want to do it. Explain that it is not personal, but rather understanding "customers like them" or "help you do/get more with...". Actually mean it or don't bother. A healthy proportion should answer the survey, and a proportion will gift you the opt-in if you are truly deserving of it - your questions were reasonable and weren't too probing. Get a proper researcher to write the questionnaire. Keep it on brand. For crying in a bucket, don't ChatGPT the thing!

You now have a single source of information about a small subgroup of your customers, with proxies. Resist the pressure to change the survey questions - regard this as a tracking study. Lock it down.

Step 4: When needing additional insights, conduct a survey amongst an online panel, targeting customers within your sector. Quota your survey according to profile of "your CRM tracking survey". In this survey re-ask the 20 data points, alongside the information you need. You are now free to ask more probing insightful questions. These questions must be shown to correlate with some of the 20 data points.

Step 5: Use the 20 common data points to fuse the online panel survey onto your CRM, which now also includes all the purchasing information you've collected since. Real basket sizes, frequency and spend. Solid and credible, not generalized - the kind the boardroom cares about. You now have a dataset which sits separate to your CRM, containing deep dive information on your actual customers alongside actual purchase behaviour, and e-mail marketing click through rates.

You also achieved this without being creepy or paying a consultant bag of cash. It might not be marketing heaven, but you're certainly halfway there, and kicking the stuffing out the competition.