Happy it's working for you. But "clients love it" is the easiest metric to game in the SaaS world.
Let's talk about what you're *really* paying for.
* The mood board feature is a classic "land and expand" wedge. Cheap entry point, but wait until you need to collaborate with more than two external users.
* Their per-seat pricing model assumes you'll onboard the entire client team. That's where the real margin is. Check your contract for "guest user" definitions and limits.
* The export options? Basic. Need high-res assets or a proprietary format? That's likely a higher-tier plan.
Used this tactic myself for years. You get praised for the shiny front-end, while the lock-in and upgrade path are built into the architecture. What's your actual user count versus paid seats?
If they offer a 'free demo', you're the product.
I'm a lead engineer at a mid-market e-commerce platform. We run our observability stack on Prometheus/Grafana/Loki, with Datadog and Splunk in use by other teams.
Core comparison for logging tools in your scenario:
**Actual log volume cost**: Datadog starts around $0.10/GB ingested, but at our volume (~500GB/day), the effective rate was closer to $0.07 with commitments. Splunk's classic license model ran us ~$4,200 per GB/day annually for ingest. OpenTelemetry to Loki/Grafana Cloud was under $2k/month for the same pipeline.
**External user collaboration**: Splunk's "role-based" access is enterprise-ready but complex. Datadog's log sharing for external clients (like a client mood board audit) is a clunky PDF export or dashboard snapshot, not real-time. Grafana's anonymous view links are better for this.
**Where it breaks**: Datadog's log explorer UI slows to a crawl on queries over 7 days at our scale. Splunk's search performance is fine but their cloud ingest pipeline choked on bursts above 2.5k events/sec. Loki with a bolted-on frontend can't match the query flexibility of either for ad-hoc searches.
**Lock-in and migration**: Splunk to OSS migration is a 6-8 month project for schema and saved searches. Datadog's agent is trivial to remove, but you lose all your dashboards and monitors unless you rebuild. Vendor engineering support was a non-starter until we hit ~$40k/month with either.
My pick is Grafana Cloud with Loki if your primary need is internal observability with occasional client sharing. If you have a dedicated SRE team and require deep, ad-hoc forensic searches across years, Splunk is still the answer. Tell us your average daily log volume and whether you need real-time client access to dashboards.
You cut off at the end there, but you're hitting on the real operational pain points: vendor lock-in and migration timelines. The six to eight month figure for a Splunk migration rings true from what I've seen in the community.
Your point about Loki's query flexibility is fair, but that's often a deliberate trade-off for the cost profile. The key is aligning the tool to the use case - using it for known debug patterns and high-cardinality event tracking, not as a general-purpose exploratory search engine. For that ad-hoc deep dive, a lot of teams are pairing it with a purpose-built tracing or profiling tool instead of trying to make logs do everything.
- GG