Here we are in mid-October, and I have a lot to share about where we are as a Technology team. We’ve wrapped up an exceptional Q3 with a fairly solid September, and we’ve now set a course forward for Q4 with a crisp, tactical plan for the current month. There’s a lot to cover so let’s get started!

How Was Q3?

Our product planning for Q3 revolved around narratives and metrics. As we continue to mature and bond as a Tech team, our first mission was to pivot from a focus on shipping deliverables (“Look, mom! I made a thing!”) to a clearer focus on the why that motivates us to initiate a technical project. The second mission was to set measurable targets and track our results.

So what were the stories for Q3? What was our why?

1. Automation starts at onboarding
2. Acquiring the right users
3. Proactive Security, not reactive security
4. Products that converse

How We Think About Narratives

In each case, we saw the stories through two separate perspectives, the perspective of our clients and the perspectives of our agent workforce. For story #1, we asked ourselves how we could provide a better onboarding experience for our clients. We saw some gaps in the Client Dashboard experience so we took those by the horns. We created interfaces for self-serve subscription changes, top-ups, and cancellations. We automated the execution of our rollover policy that allows clients to spread out their use of purchased credit. We implemented support for teams to use our services rather than assuming that our clients engage with us one at a time. An aggressive set of goals, and we delivered!

We also asked ourselves how to smooth the onboarding experience for our agents. We believe a happy workforce is critical to a happy customer. Every service business has this exposure — if your team is unhappy, how can they possibly have a positive demeanor when they engage with your client? The biggest gap we saw in this space was an overly ad hoc set of processes for compensation. While agent compensation was already largely automated, we found that there were a few gaps that created dependencies on a few overworked partners. So we planned and ultimately delivered a major refactor of our compensation system for agents called Agent Pay 3.0.

How We Measure Outcomes

So how do we measure these programs and understand success or failure? In each of the cases above, our primary metrics were related to efficiency. How many hours are spent approving agent payments? Let’s reduce it by 50%. How many hours are spent manually processing subscription changes for clients? Let’s reduce that by 80% — ideally we’d kill that one altogether!

Some initiatives lend themselves more to measurable outcomes and ROI, but the key here was to establish a practice of measuring. Note that these metrics didn’t yet tie off to overall company objectives. They were also largely directional and focused on what we thought we could achieve rather than what we should.

And How Did We Do?

At the highest level, we did pretty darn well. We didn’t blindly chase results. We dropped a few projects when they stopped making sense and added a few that needed to move up in the timeline.

https://share.getcloudapp.com/o0uxjA7R

Major Deployments in Q3

We had a lot of headlines in Q3. Here are a few of the bigger ones:

• Agent Pay 3.0 drastically reduces manual effort associated w/ agent compensation
• Automatic Routing PoC shows that we’re ready to embed machine learning in our platform
• Automated Billing interfaces allow clients to self-serve as well as reducing internal costs
• Limited Rollover enforces our terms of service and improves margins
• Process Efficiency 1.0 introduces quoted prices and enables efficiency gains
• Team Management 1.0 allows multiple users in a company to share access to our services
• Automated Messaging foundations laid with Segment and tools for custom messages
• Client Onboarding improvements ensure better fit clients
• Agent Profile and Client Profile in DAL enable better context for operators
• Process Builder 2.0 introduces the notion of tasks (small process steps) to the DAL

What About Other Objectives?

Not all our goals fit into the category of product deliverables. We set separate goals for Engineering related to reliability, timeliness, and innovation. That included a zero tolerance policy for full service outages, a commitment to improved burndowns per sprint, and improvements to our underlying approach to data.

We underachieved a little bit on reliability by releasing an update to our internal tools that caused an outage for our Ops team. We were also hit by outages from a few cloud providers. Our timeliness was quite good overall, and this has radically improved our ability to forecast how much work we can plan for upcoming sprints. We introduced better practices in the data space by implementing a Kafka solution where we produce immutable fact streams related to how we execute on client delegations. This is extremely exciting progress that sets the stage for more maturity in our machine learning story in Q4!

We also managed to expedite our deployment of virtual desktops for the agent workforce. This improves our security positioning as it gives us a vise grip on enforcement of a standard desktop for agents. We piloted Amazon Workspaces late in Q2, and we made it more production-ready in Q3. We’ll be deploying more broadly across the workforce in Q4.

One overall, extremely subjective observation — we focused too much on timely deliverables and incurred some architectural debt. This is generally the result of well-intentioned engineers trying to hit timelines and help the rest of the company as quickly as possible. We always need to give Engineering room to make good long-term decisions. And we should always remind them that it’s desirable to do so.

And How Was September?

Not our best month — more details here. We closed out the month a little shaky on key deliverables after a stellar beginning. This was largely due to some complex work related to landing Quotes and Tasks as core concepts in the DAL. Happily, those are now in place today.

The Plan for Q4

How can we do better in how we approach roadmapping for Q4? We made great progress in moving beyond a mere project plan, but we ran into some limitations. Are our narratives the most important “why’s” for the company? Are the metrics arbitrary or are they well-targeted? And how do they connect to the financial goals for the company overall?

So this is how we’re getting better. Instead of narratives and metrics, we have narratives with metrics. And they connect to our high-level company goals, which are:

• Grow the business / Drive MRR to $160,000 a month
• Expand usage with current clients / Reduce churn to <5% of MRR
• Drive to profitability / Increase gross margins from 30% to 50%
• Deliver a first-rate service / Improve NPS (still measuring current state)

I’ll be publishing our full roadmap separately, but here are the high level OKRs that motivate us.

• Increase gross margins by 25% by reducing direct labor COGS.
• Create a safe and secure client experience so that 80% of delegations require no manual handover of credentials.
• Reduce overhead per delegation by 25% through DAL automations.
• Drive engagement — no client will use less than 85% of their hours per month.

More to come soon!