Okay, I'll go first since I started this thread. We're evaluating a new vendor after our contract renewal came in with a 40% hike. I can't name names due to NDA, but I can share the structure and ballpark figures, which I think is more useful anyway.
Our setup: multi-touch attribution (MTA) platform, ingesting ~1.2 billion marketing events monthly via a mix of S2S connectors, cloud storage (S3/GCS), and a Kafka topic for real-time web events. We have 18 months of look-back, and we use their cross-device graph (deterministic + probabilistic).
**Our current costs (annual contract):**
* **Base Platform Fee:** ~$220k
* **Event Volume Overage:** Tiered pricing kicks in past 1B events/month. We typically hit ~$15-30k in overage charges quarterly.
* **Data Onboarding/Connector Fee:** One-time $25k for custom engineering to handle our unique streaming event schema (this was a fight).
* **Cross-Device Graph Access:** Bundled, but they quoted $60k as a standalone line item.
All in, we're looking at **~$350k-$400k annually**, and that's before any professional services for model tuning.
What I'm really curious about: how much of your cost is tied to raw event volume vs. "seat" licenses for analysts? Are you seeing more vendors move to a cloud-consumption model (like cost per million events processed)? And does anyone have a line item for "cookieless measurement" capabilities yet, or is that still baked in?
Hoping to see some numbers to understand if we're in the right ballpark or getting fleeced. The data pipeline costs to feed these things are significant on our end too, so the total cost of ownership is wild.
That's a useful baseline, especially the detail on the custom connector fee. It's a common cost sink that often gets buried in initial conversations.
In my experience, the seat/license cost becomes significant primarily when you have a large analyst team needing concurrent access to the UI for deep exploration and model building. If your primary use case is automated reporting feeding into a BI layer, you might only need a handful of seats. The vendor you're describing seems to be pricing heavily on infrastructure burden (event volume, look-back window, graph computation), which aligns with their >$1B event scale.
When we last negotiated, we pushed hard to decouple seat costs from the core platform fee. We ended up with a base package covering infrastructure and processing, with analyst seats as a separate, scalable add-on. This stopped us from paying for 50 'potential' users when only 8 ever logged in.
Your overage charges at that volume are interesting. Have they been consistent, or do they spike with specific campaign cycles? That tiered pricing can create unpredictable quarterly budgeting.
Your structure aligns with what I've seen. The base fee at ~$220k for 1B events/month is a common anchor, but the overage charges you mention can spiral. I'd scrutinize the tiered pricing model closely.
>how much of your cost is tied to raw event volume vs. "seat" lic
From my benchmarks, at your event volume, over 70% of the cost is usually tied to infrastructure: event ingestion, look-back storage, and graph computation. Seat licenses are often a secondary line item, maybe 10-15% of total cost, unless you have a huge team of analysts constantly in the UI building new models.
A key negotiation tactic is to cap the overage fees annually. We pushed for a hard ceiling at 1.5x our committed volume, which saved us nearly $40k last year when we had a traffic surge. Without that, the quarterly overage becomes unpredictable. Have you tried negotiating a true-up model instead of quarterly overage charges?
Totally agree on decoupling the seat costs! That saved us so much heartache during our last renewal. We managed to structure it as a core 'engine' fee for processing, and then we buy analyst packs of 5 seats at a time. It's way more predictable.
Your point about overage spikes is so key. We absolutely see them tied to big campaign cycles, especially around Q4 holiday pushes or new product launches. It makes budgeting a nightmare. We actually negotiated a "campaign surge buffer" where we get an extra 20% event volume allowance for two specific months of the year without overage penalties. Maybe that's something you could push for next time?
test everything twice