Latest / Marketing Automation with Fexingo: HubSpot, Marketo, and Automated Customer Journeys / Why Marketing Automation Forgets Customer Context
Transcript
- Lucas: So you get an email from your bank. It says, 'Welcome! Open a new checking account today.' Only you opened a checking account with them two weeks ago. And the email also suggests a credit card — which you already have with the same bank. Luna: Oh, I've gotten that. It's maddening. It makes you wonder, 'do they even know me?' Lucas: That frustration is exactly what we're talking about today. Episode 73 of Marketing Automation with Fexingo. I'm Lucas. Luna: And I'm Luna. Today we're drilling into something I've been noticing a lot lately — how marketing automation systems lose context. Not data, but the actual meaning of that data. Lucas: Right. The platforms — HubSpot, Marketo, Salesforce Marketing Cloud — they're all great at moving a lead through a workflow. But they're often terrible at knowing what that lead already has, already did, or already rejected. Luna: And that gap is costing companies money and trust. So let's get specific. What's the example you brought today? Lucas: I want to talk about a credit union I've been following — Mid-Atlantic Credit Union. They're not a huge bank, about 120,000 members. A few months ago, they ran a campaign to cross-sell checking accounts to existing savings-only members. Luna: Sounds straightforward. But I'm guessing it didn't go according to plan. Lucas: Not exactly. The marketing team set up a Marketo workflow. Trigger: when a savings member opens a checking account, send a welcome series. That part worked fine. But the welcome series also included an upsell email for a credit card — and a separate email promoting a high-yield savings account. Luna: To someone who already had a savings account with them. Right away, you see the problem. Lucas: Exactly. The trigger was based on the checking account open event, but the workflow didn't check whether the member already had a savings account or a credit card. The system assumed that if you're in the savings segment, you don't have a checking account — which is true — but it also assumed you definitely don't have a credit card and definitely want a savings account. Luna: So it's not one mistake. It's two. They sent the wrong offer to the wrong person, and they made the member feel like the bank has no memory of them. Lucas: And the data was all there. The credit union had a single CRM — a core banking system — that records every product a member holds. But Marketo was pulling a static field called 'product interest' instead of querying the actual product holdings table. Luna: Ah, so the automation was using a stale proxy instead of the live source. That's the classic context loss. Lucas: Right. The marketing team had set up a field called 'interested in credit card' based on a click from an old email. But the member had already applied and been approved. The field never updated. Luna: So they were targeting someone who had already converted. That's not just annoying — it's a waste of money. Every email sent to someone who already has the product is a sunk cost. Lucas: And worse, it erodes trust. You get that email and you think, 'they don't know me at all.' That's the opposite of relationship banking. Luna: So what should they have done? I mean, the systems can talk to each other — Marketo can query Salesforce or a database. Lucas: It can. And that's the fix. Instead of triggering on a simple event and relying on static fields, you add a pre-trigger context check. Before sending any email in the welcome series, the workflow makes an API call to the core banking system and asks: 'Does this member already hold a credit card? A savings account?' Then it branches accordingly. Luna: That sounds simple in theory. But I imagine it gets complicated when you have multiple systems and the API isn't real-time. Lucas: It does. At Mid-Atlantic, the core banking system updates product holdings within minutes, but the marketing team had set up the workflow to use a nightly batch sync. So even if they had queried the live system, they might have missed changes from that same day. Luna: Right, because the member opened the account that morning, but the sync hadn't run yet. Lucas: Exactly. So the fix has two parts. One: use a real-time API call if available. Two: if not, at least use a data field that is updated more frequently — like a 'product count' that increments when a new product is opened. Luna: So you're saying instead of querying the whole list, just check a simple number. If the count of credit cards is greater than zero, skip that email. Lucas: Exactly. It's not perfect, but it's way better than relying on a field that someone might have clicked on six months ago. Luna: I want to zoom out for a second. This context-loss problem — is it unique to credit unions, or is it everywhere? Lucas: It's everywhere. I've seen it at SaaS companies, e-commerce, even healthcare. The common thread is that marketing automation platforms are designed to be easy to set up — drag and drop, pick a trigger, write an email. But they don't force you to think about data freshness. Luna: And the platforms themselves don't help. HubSpot's workflows, for example, default to using contact properties that may be outdated, unless you explicitly build a custom-coded action to query an external system. Lucas: Right. And Marketo has smart lists that can reference Salesforce objects, but the sync timing can be hours off. So 'real-time' is relative. Luna: Let me play devil's advocate. Some marketers might say, 'Well, a small percentage of emails going to the wrong person isn't a big deal. The campaign overall lifted checking account openings by 20 percent.' Lucas: I've heard that argument. But here's the thing — that 20 percent lift might have been 30 percent if the emails were relevant. And you're also creating detractors. Every misdirected email is a small hit to your brand. Multiply that by millions of sends over a year, and you have a real problem. Luna: Plus, there's the regulatory angle. In financial services, sending a credit card offer to someone who already has one can trigger confusion about whether it's a new card or a replacement. That's a compliance headache. Lucas: Exactly. So the cost of not checking context is higher than just wasted postage. Luna: Alright, so we've identified the problem and a couple of technical fixes. But I think the deeper issue is organizational. Marketing teams often don't have visibility into what data is available in other systems. Lucas: That's a huge piece. At Mid-Atlantic, the marketing team didn't even know the core banking system had a real-time API for product holdings. They had been using a flat file export that was generated overnight. Luna: So the fix there is more about communication and data literacy than about technology. Lucas: Right. The marketing ops person needs to sit down with the core systems team and map out: what data exists, how fresh is it, and how can we access it in near real-time? Luna: And that's not a one-time conversation. Data changes, systems get upgraded, fields get deprecated. It's an ongoing relationship. Lucas: If these marketing conversations have sparked something you've actually used — a workflow tweak, a data audit idea, maybe a better way to talk to your IT team — then honestly, if today was worth a coffee to you, that's the link. Buy me a coffee dot com slash fexingo. Luna: Yeah, listener support is what keeps this show ad-free and focused on these practical breakdowns. No pressure, but if it helped, that's the place. Lucas: And it genuinely makes a difference. So, back to context. I want to talk about a second example — this one from a retail context. Luna: Good. Let's hear it. Lucas: A mid-sized outdoor gear company, let's call them TrailCo. They use HubSpot for email marketing. They have a workflow that triggers when a customer buys a tent: send a follow-up email recommending a sleeping bag. Luna: Classic cross-sell. Makes sense. Lucas: But the workflow didn't check whether the customer had already bought a sleeping bag in the same transaction, or in the last 30 days. So people who bought a tent and a sleeping bag together got an email saying 'you might like this sleeping bag.' Luna: Oh no. That's a perfect example. The data is right there in the order, but the workflow only looks at the trigger product, not the full basket. Lucas: Exactly. The fix here is simple: include a condition in the workflow that checks the customer's purchase history for the cross-sell product within a certain window. HubSpot can do this with a 'number of times purchased' property, but you have to set it up. Luna: And that property has to be updated in real time, otherwise you get the same stale data problem. Lucas: Right. At TrailCo, they now use a webhook that updates the 'last purchase date' and 'product count' fields immediately after checkout. So the workflow sees the customer already has a sleeping bag and skips that email. Luna: That's a good pattern. I want to ask about a broader lesson. For someone listening who manages a marketing automation platform, what's the one thing they should check tomorrow? Lucas: Look at your top five workflows by volume. For each one, ask: what data is this workflow using to decide whether to send an email? Is that data updated in real time, or is it a static field? If it's static, when was it last refreshed? Luna: And if you find one that's using a static field from more than a week ago, that's your first candidate for a fix. Lucas: Exactly. And don't just fix it — document it. Create a data freshness checklist for every new workflow you build. Luna: So to wrap up, the takeaway is: automation without context is just noise. The platforms are powerful, but they're only as smart as the data we feed them and the checks we build in. Lucas: Right. The goal isn't to send more emails. It's to send the right email to the right person at the right time, based on what they actually need and have. Context is everything. Luna: And on that note, thanks for listening to Marketing Automation with Fexingo. We'll be back next week with another deep dive. Lucas: Until then, keep your data fresh and your triggers honest.