Does the mention of a Customer Data Platform (CDP) bring on collective groans in your company?
Maybe you’ve been scarred from a legacy CDP that burnt marketing budget and took years to untangle from. Or maybe you’ve been pitched a ‘new wave’ CDP that over-promises but still under-delivers.
Either way, the concept of a CDP doesn’t seem to address the primary challenge your brand has with customer data.
You DO need to feed your marketing channels with identified customer event data, but you DON’T need to maintain a complex data architecture to do that.
I believe stores who’ve made the enlightened decision to move their ecommerce operations to Shopify can also do away with much of the cost and complexity of maintaining a ‘modern data stack’ - while keeping the same benefits.
In fact, Shopify stores can take this opportunity to further boost their marketing performance by replacing legacy pixel-based tracking with server-side tracking, and reusing profile identification across the main marketing channels.
The CDP promise
Let’s first dissect what a CDP is supposed to do; and then look at which parts can be simplified for a Shopify brand.
The main ten components of a CDP are:
1. Web collection – adding a script or pixel to the storefront to track customer interactions
2. Cloud data sources – pulling data about customers, orders and products which are collected in other customer tools (e.g. CRM)
3. Schema validation – filtering events to ensure a consistent data schema
4. Profile building - merging the customer data points into a unified customer profile, including handling different customer identifiers
5. Segmentation - building segments of customers based on behavior
6. Predictions – predicting customer lifetime value (LTV) and other predicted traits based on similar customer behavior
7. Transformation – ensuring event data is in the right format, with the right customer traits, for the data destination to accept
8. Cloud destinations – streaming events into other marketing, analytics and customer platforms
9. Warehouse sync – pushing events into a data warehouse
10. Pipeline monitoring - checking all of the above is working seamlessly as expected
Sounds like a lot to set up and maintain? Absolutely! I’ve seen mid-market DTC brands waste 6 figures just with the CDP set up, without starting to get value from the resulting unified profiles.
The simpler alternative
Brands on Shopify can simplify this data pipeline drastically: there are typically just two sources – the store frontend and Shopify’s servers – and half a dozen marketing data destinations.
Looking at it step-by-step you can see where we can eliminate half the complexity:
1. Web collection – still required. Maybe the most critical part of the stack; garbage in, garbage out.
2. Cloud data sources – required but simpler. Almost all customer data is synced back to Shopify, so a real-time connection with Shopify covers most needs
3. Schema validation – not required. Standard ecommerce events to capture a standard shopping journey match a standard ecommerce schema
4. Profile building – required. Especially the unification of anonymous (cookie IDs) and identified (by email, phone) profiles.
5. Segmentation – not required. This is done in the downstream destinations. E.g. Klaviyo segments can be synced with Meta directly.
6. Predictions – not required. Also down in downstream destinations.
7. Transformation – required but simpler. Unified ecommerce schema into standard destinations = automatic transformation.
8. Cloud destinations – required but simpler. DTC brands are all using the same marketing tools (Google, Meta, TikTok, Klaviyo etc)
9. Warehouse sync – not required. What are you using that warehouse for except for transformation, segmentation and predictions? Data warehouses are for huge enterprises with hundreds of data sources.
10. Pipeline monitoring - required but simpler. Automation means there is less to go wrong.
With all this simplification, to achieve the same end – efficient and accurate engagement of customers and prospects – we can drop the cost of ownership of the data pipeline 10x.
Take a DTC brand with $20M GMV, getting millions of store visits monthly. Under a Monthly Tracked Users pricing model, they would be paying $100k+ annually for CDP software, and the same again for an agency or team member to manage the setup. That’s a $200k investment before you get any tangible results.
By renting Littledata’s data pipeline for Shopify they would be paying less than $20k annually to cover the same needs.
What if you chose to go headless Shopify with Nacelle? Littledata can still support you. Our server-side tracking has been built for headless or custom storefronts, linking together the marketing campaigns, with the pre-checkout journey, with the Shopify checkout.
Data warehouse vs data store
So maybe I was too harsh on the data warehouse part of this stack.
There are still a few valid reasons why you want to aggregate all your customer data on a server that you control.
Firstly, it stops you being held random when switching tools. Want to switch email providers and get the same targeting potential? You’ll need to relay the same customer events to the new provider.
Secondly, it provides a way to train your own AI models on your proprietary customer data set. AI agents are only going to get better at understanding structured data sets – especially if they use a consistent schema.
What I object to the term ‘warehouse’: it sounds like a big dusty shed where trucks from all over bring in pallets of data, which then get stacked and never used.
Instead, I prefer to think of a data store – where one big table of customer events with a consistent schema, all unified into profiles, can be exported to any other tool that needs it. It’s your data insurance policy. It doesn’t need to sit on your own infrastructure – but it does need to be fast and easy to access.
Klaviyo’s Data Platform offering is half right on this point. Klaviyo wants to provide brands with access to the identified customer profiles as stored on Klaviyo for use in other marketing tools … but this does not include all the anonymous events from prospective customers. Klaviyo is blind to what happens before a customer signs up for email – which is 90% of visitors on the average Shopify store.
For now Littledata offers a robust connector for Shopify to Twilio Segment to fulfil this need to store both the anonymous event stream (for mid-funnel retargeting using ads) and the identified customer profiles.
And later in 2025 we’ll enable real-time event streams and batch data export APIs to pipe the customer data wherever you need it.
That makes it a great fit for tools like Paige, Nacelle's AI, to assist with Progressive Identification of the anonymous visitors.
Final thoughts
You don’t need a CDP - and you don’t need data engineers. But you do need unified customer data.
The term ‘CDP’ gets associated with an expensive, monolithic legacy tool; but accurate customer data doesn’t have to be that way.
To set your customer data free in the marketing tools you choose you need to first lock down the event data schema. Then you can reap the benefits of simplicity.
Learn why Shopify brands don't need expensive CDPs to unify customer data. Discover how to simplify your data pipeline while boosting marketing performance