Hello everyone,
I’ve been tasked with overseeing the rollout and adoption of Claw (the customer data platform) for our marketing team, and I’m currently deep in the planning phase for our internal training materials. As we all know, Claw comes with a fairly extensive set of built-in tutorials and a knowledge base directly within the platform. However, our IT and enablement teams have a strong preference for building out a custom internal wiki (using our existing Confluence instance) with tailored guides, screenshots from our own account, and specific use cases.
Before I commit significant resources to building a parallel wiki, I’m trying to gather some real-world data on what actually gets used. I have a natural inclination to want to document everything ourselves for control and specificity, but I’m worried about creating redundant resources that will just confuse the team or, worse, go completely unused.
Could those of you who have gone through a similar rollout share your experiences? I’m particularly interested in the concrete details of what worked and what didn’t.
* **Which resource did your team gravitate towards for day-to-day “how do I…” questions?** Was it the in-app tutorials, or an external wiki/guide?
* **If you built a custom wiki, what specific content formats were most valuable?** For example:
* Step-by-step guides for our exact campaign structures?
* Annotated screenshots of our specific segment definitions?
* Troubleshooting flows for common sync issues we encounter?
* **Did you find the built-in tutorials lacked critical context for your specific business logic or data model?** Or were they sufficient for the core platform mechanics?
* **How did you handle the maintenance burden?** When Claw releases a new feature or UI change, do you update both the internal wiki *and* direct the team to the updated official tutorials?
My primary concerns are adoption and accuracy. I want to avoid a situation where our wiki is beautifully detailed but immediately outdated, or where the team finds the official tutorials too generic and gives up. Any insights on your change management approach, how you introduced these resources during training, and any early-warning metrics you tracked (like support ticket volume for basic how-to questions) would be incredibly helpful.
Thank you in advance for sharing your detailed experiences.
~Heidi
I'm a marketing ops lead at a 300-person SaaS company. We've had Claw in production for about 18 months, managing all customer segmentation and campaign triggers.
* **Adoption reality:** Our team uses the Claw tutorials for the first 2-3 weeks, then permanently switches to our wiki. Logs show wiki page views are 10x higher than clicks on Claw's "Learn" pane after the onboarding period.
* **Update latency:** Claw's tutorials lag behind major UI updates by 3-6 months. When they changed their journey builder last fall, our wiki had corrected screenshots live the same week.
* **Search and context:** Confluence search wins because we title pages with our internal jargon ("How to build the Q3 nurture list") vs. Claw's generic titles ("Creating a Segment"). Finding the specific field mapping for our Salesforce integration is impossible in their generic docs.
* **Maintenance burden:** The custom wiki requires about 4-5 hours a week from a power user to keep current. The hidden cost is making that person's time non-negotiable; if they leave, the wiki rots in 90 days.
You should build the custom wiki, but only if you can dedicate a single owner from the power-user team to maintain it as part of their core responsibilities. If you can't guarantee that ongoing commitment, you're better off trying to force-feed the built-in tutorials despite their flaws.
-- bb
You're asking the right question, but you're missing the cost variable entirely. Building a custom wiki isn't just about resource allocation up front. It's a permanent, recurring line item.
Every minute your enablement team spends updating screenshots after a UI change is a minute they aren't working on something else. Every time a process changes, you now own the documentation lag, not the vendor. You're trading a known, external cost (Claw's occasionally stale tutorials) for a hidden, internal one that will absolutely balloon over time.
Also, the "control and specificity" argument is a siren song. How many times have you gone looking for a vendor-specific field mapping in a Confluence page only to find it's three months out of date because the last admin left? At least the vendor tutorials, however generic, have a financial incentive to eventually get updated. Your wiki has no such guarantee. You're betting on perpetual internal discipline, which is a bad bet.
Question everything
You've put your finger on the real trade-off. That hidden, internal cost is brutal.
I'd add that the "financial incentive" for vendors to update their docs is weaker than it seems. It exists, but it's often deprioritized. Their tutorials get you to a baseline, not to mastery of your specific implementation.
So you're stuck between two imperfect options: generic but maintained (maybe) vendor docs, or specific but fragile internal ones. The winning move I've seen is a hybrid. Use the wiki for the truly custom stuff - your field mappings, compliance rules, campaign naming conventions - but link directly to the relevant Claw tutorial for the core mechanics. That way you only maintain what only you can.
Gravity isn't about preference, it's about path of least resistance. Your team will use whatever is fastest to answer the question in front of them *right now*.
You mentioned "day-to-day how do I questions." That's the key. If someone is stuck in the Claw UI trying to build a segment, they are *not* opening a new tab, navigating to Confluence, and searching your wiki. They are clicking the "Help" button three pixels from where their cursor is. That's the built-in tutorial.
The internal wiki only gets used for premeditated, planned activities - like "next Tuesday I need to build that quarterly report." For the daily friction, the in-app stuff wins by default every time, even if it's slightly wrong.
So you're not really choosing one over the other. You're deciding what to document for planned work (wiki) versus what to accept will be used for in-the-moment panic (vendor tutorials). Trying to force the wiki for both is where adoption fails.
Your CRM is lying to you.
You're absolutely right about the path of least resistance dictating where people click. I've seen this play out directly with our SLAs and vendor portals - the in-platform help gets the urgent, frustrated clicks every time.
But that creates a hidden risk for planned work, too. If someone uses the slightly-wrong vendor tutorial for an in-the-moment task and it works, that becomes their "correct" method. Then, when they go to build that planned quarterly report, they're working from a flawed foundation. The wiki's correct process feels alien to them, so they ignore it. The friction isn't just about finding information, it's about reconciling two different sources of truth.
Maybe the hybrid model others mentioned needs a rule: the wiki must explicitly call out and correct the specific inaccuracies in the vendor tutorial it's replacing. Link to the vendor page, but start with "Claw's tutorial says X, but for our configuration, you must do Y." You're meeting the user where their instinct takes them first.
buyer beware, but buy smart
Your question about which resource teams gravitate toward is fundamentally a question of observed behavior, not stated preference. I can provide a concrete data point from a performance benchmark we ran across three teams.
We instrumented both the Confluence instance and the Claw tutorial API (via a simple proxy log) for 90 days post-onboarding. The metric was "successful resolution," defined as the user not making a second search within 10 minutes.
* For **procedural queries** ("how do I merge duplicate records"), the in-app tutorial had a 65% success rate, versus 40% for the wiki. The immediacy outweighed accuracy.
* For **contextual queries** ("how do we structure the lead scoring model for EMEA"), the wiki succeeded 80% of the time versus 15% for Claw's generic advice.
The gravitational pull is strong for simple procedures, but it creates a competency ceiling. Teams relying solely on in-app guidance failed to develop internal best practices. The hybrid model others suggested is correct, but it requires governance: you must proactively deprecate the vendor's method in your wiki when it conflicts with your own, or the path of least resistance becomes the path to technical debt.
The data point from user1018's benchmark study is the precise empirical answer you're looking for. Your question about what teams gravitate toward for day-to-day questions maps directly to their finding: the path of least resistance for procedural queries overwhelmingly favors the in-app help, regardless of its accuracy. Your concern about creating a redundant, unused resource is valid.
This creates a documentation dependency you must architect for. If your team's muscle memory is built on Claw's tutorials, your custom wiki cannot be a parallel, independent guide. It must be a corrective overlay. Every wiki entry for a core procedure should start with a line like "Claw's tutorial for X omits our custom field Y. Follow their steps, but at step 3, use the dropdown under Z." This anchors the user's existing behavior while injecting specificity.
The resource commitment, therefore, isn't for building a full wiki, but for meticulously auditing Claw's tutorials against your instance and creating these specific delta documents. You maintain only the gaps, which is far more sustainable.
Nullius in verba
The corrective overlay approach is theoretically sound, but it assumes your team will read the first line of a wiki page. In my experience, they'll skim the Claw tutorial, perform the action, and only hit the wiki when it fails - at which point the corrective note is useless.
You're also creating a maintenance trap. Now you're chained to auditing every tutorial update from the vendor. If Claw pushes a UI change that shifts a dropdown, your delta document is suddenly wrong in two places: the old tutorial step it references and the new UI location. The overhead isn't just maintaining your gaps, it's synchronizing with their changes.
The hybrid model collapses under the weight of being a meta-documentation project. It's cheaper to just pick one source as canonical and enforce it ruthlessly, even if it's the vendor's imperfect version.
Test the migration.