By Jane Steen, Head of Engagement at ETL
Most sales and marketing teams are not short of data. CRM records, campaign analytics, web sessions, transaction history, feedback forms…all of it accumulating, and much of it underused. The familiar story is that teams need more data, newer tools or a dose of AI. Our experience is usually the opposite. The data is already there. It’s just scattered across systems, missing context and rarely joined up in ways that answer the question in front of you.
This does change what “doing better with your data” looks like in practice. Two disciplines in particular tend to earn their keep if the data ecosystem is complete.
The trouble with customer feedback isn’t the responses
According to HubSpot, 42% of companies don’t collect customer feedback at all. That figure surprises people, although it probably shouldn’t. Feedback programmes are time-consuming and their ROI can be awkward to pin down, so they sometimes stall after the first survey wave.
Even where feedback does get collected, the limiting factor is rarely the responses themselves. It’s the context around them. A 3/10 score from a new customer on their first order means something quite different from a 3/10 from a ten-year account who had a delivery go wrong last week. Without the context you have a number. With it, you have something a marketing team can actually act on.
Useful feedback data sits in layers. There’s the customer itself: who they are, what they’ve bought, how long they’ve been with you. There’s the touchpoint: what happened, where, through which channel and which team. And there’s the context around that interaction: time since last contact, any recent issues, whether anything is still open. The response sits on top of all of that.
Pulling those layers together is where most of the effort goes, and it’s usually the part that gets skipped. We often see feedback programmes that capture answers cleanly but can’t reconnect them to the customer’s history, because the history lives in three different systems with three different identifiers. Fixing it is rarely glamorous. It’s mapping fields, reconciling records, making sure consent flags are in the right place for GDPR purposes and checking the joins hold up at volume.
None of that shows well in a deck. It does turn feedback from a scorecard into something the business can use.
Customer lifetime value, done with fewer shortcuts
Customer lifetime value (CLTV) has been around long enough that most teams think they already know what it is. In the simplest form, CLTV is customer value multiplied by customer lifespan, where customer value is average purchase value multiplied by average purchase frequency. Straightforward on paper.
In practice, CLTV is one of those metrics that is easy to calculate and easy to mislead yourself with. A few patterns come up repeatedly.
The first is running a single CLTV figure across the entire customer base. High-value and low-value customers pull the average in opposite directions, and the number you end up with is usually too generic to drive a decision. Segmenting by acquisition channel, product line or cohort tends to tell a more honest story. It’s sometimes an uncomfortable one about which channels look cheap but deliver customers who don’t stick around.
The second is lifespan. CLTV calculations lean on an assumed retention rate, and that figure is harder to pin down than it looks. Retention varies by cohort, by acquisition channel and over time, and simple averages tend to flatter the number rather than challenge it. Overstated lifespan inflates CLTV, and the acquisition cost that looks affordable against the inflated figure often isn't.
The third is treating CLTV as a one-off calculation. Customer behaviour, pricing and retention all shift. A figure calculated eighteen months ago is a historical artefact, not a benchmark. Recalculating quarterly is a reasonable baseline; for faster-moving categories, more often than that.
The fourth, and the one we run into most, is data silos. Capturing a full customer journey is close to impossible when transaction data lives in one system, service interactions in another and marketing engagement in a third. The CLTV that gets calculated in that environment isn’t wrong exactly. Instead, it’s often a view of whichever system happened to be easiest to pull from.
The principle extends beyond customers, too. One of our automotive clients applies a similar calculation to individual vehicles, tracking which deliver the most profitable lifespan across sale, servicing and parts. It’s the same discipline, just a different unit of value. It allows our client to focus marketing and aftersales investment where the long-run return actually is, rather than where last quarter’s revenue came from.
What the two have in common
Neither customer feedback programmes nor CLTV models are new ideas. But both are still where sales and marketing teams find measurable lift when the data is provided carefully.
The reason is that they reward the unglamorous data work, such as identifiers, consent flags and customer record reconciliation. They reward teams who agree a definition of what a “repeat purchase” actually is. None of it sits well on a roadmap slide, but all of it determines whether the roadmap works.
For teams deciding where to put the next bit of data effort, the more useful question is rarely “what’s the newest tool we should pilot?” It’s closer to: which of the decisions we’re making today are being made on incomplete data, and what would it take to fix the join?
The answer is almost never as interesting as a new platform launch. But it does tend to be the one that pays for itself.
ETL is a Manchester Digital member. We work with sales, marketing, operations and IT teams to get customer data out of silos and into a form people can use. If anything in this article sounds familiar, we’re happy to compare notes.
