Sometimes it's better to build than buy
Welcome to Beyond The Click by Balboa Solutions. In today's issue we're sharing the current state of Product in two stories - one micro, one macro:
Sometimes it's better to build than buy
AI changes the revenue retention game
👉 Real quick:
Ever wonder if your in-app guides actually teach your users how to be smarter users? Or if they just click through and forget? On June 12 we're talking about ALT (Adult Learning Theory) and how it can improve your user enablement.
OK, let's dive in.
1. Sometimes it's better to build than buy
AKA: How to build a successful internal-use app for your enterprise 🏦
As mature as the SaaS industry has become, including...
the public cloud
industry-wide best practices in customizability and user-friendliness
the digital upskilling of knowledge workers everywhere
...off-the-shelf solutions will never effectively cover all enterprise use cases. It's kind of like the 80/20 rule - a relatively small number of commercial SaaS products can meet the needs of most teams and processes.
But there's a lot of edge cases - workflows so unique and core to a company's competitive advantage that a solution made for the mass market, almost by definition, can't enable it. But many enterprises have the internal resources to build something that can. Some key examples:
Home Depot's store associate-optimizing Sidekick mobile app 🔨
UPS and their ORION delivery route optimization system 🚚
Delta Airlines' Flight Weather Viewer to help pilots navigate turbulence ✈️
JLL Technologies and their purpose-built real estate BI app, Azara 🏢
That last one was the subject of our latest webinar, From Feedback to Impact: Scaling Employee Experience with Data (full recording here🔗). We spoke with Ben Herkenhoff, who shared his journey with the enterprise app playbook as Director of Product Management for Azara. It's a fun story, and it changed our whole perspective on product ops for the enterprise. 👇
For starters - why even bother? 🔍
JLL Technologies (JLLT) supports JLL's commercial real estate management and advisory services. The North Star for this initiative was to compare real estate KPIs “apples to apples” across the whole business. Ben’s team took inventory of the non-negotiables the solution would require:
A single data model, enforced company-wide
A real estate-specific UX
Feature prioritization for all JLL’s current (known) and future (unknown) needs
Despite all the 3rd-party BI tools on the market, none of them could deliver all three requirements. So, Ben’s team set out to build, launch, and manage Azara - the higher-investment option, but one that would get JLL a 99% solution instead of 80%.
Launching the MVP 🚀
Ben's team resisted the urge to "boil the ocean," initially launching Azara on just one of JLL's datasets - this allowed them to prove ROI faster. Some critical first features:
Single source of truth, with consistent metrics across 80+ countries
Configurable nomenclature layer (e.g. the “Work Order” KPI on JLL's side automatically shows up as “Service Ticket” client-side)
Purpose-built for JLL employees, the primary users (e.g. facility managers, lease administrators); client users are important but secondary
Site-specific dashboards (but baseline views are still enforced)
Iterating toward v2 📈
The team embedded Pendo into Azara from launch. This included in-app release notes (much higher open rate than email blasts) and a Resource Center. It also involved feature use tracking and a 3-tiered survey cadence:
Day 0: Onboarding pulse
Day 90: Value check-in
Every 6 months after: Keep asking "How well is Azara meeting your needs?"
The team also leveraged Pendo's Feedback module with a “have an idea?” button right in the product header. Their MO has been to frequently seek user feedback and then rigorously deliver on it.
A combo of PM triage and user votes assigns feature priority in the roadmap.
As the team delivers on the request, Pendo auto-emails status updates to the submitters AND the voters.
Pendo auto-emails a “did this solve it?” survey after the change goes live.
The above process reduces inbound emails from users looking for updates on their requests. The team also started an “#azara-wins” Teams channel to celebrate positive NPS reviews, motivate the product and dev teams, and find user candidates for beta advisors.
Gaining trust ⚖️
For internal-use software, your colleagues are your clients. Ben and his team have made sure to treat them as such, including:
Letting stakeholders vent during early discovery meetings (to uncover and really hear their pain points with previous BI solutions)
Presenting a visual map of all the stakeholders they’ve consulted (so nobody fears their voice will go unheard)
Framing every launch as just an MVP and NOT the final product (and committing to make changes after watching the usage data for a bit)
In update meetings, specifically rehashing stakeholder concerns from past meetings and how they were addressed in the update (to build trust in future release cycles)
Gaining trust was the steepest learning curve. It comes down to change management and stakeholder engagement. Ben emphasized that these are extra nuanced with internal users because everyone is just trying to do their job. To get their buy-in, you have to demonstrate that your product help that instead of hurt it.
Moral of the story 👍
Best-in-class enterprises are going to keep investing in best-in-class systems. Running product ops for an internal-use enterprise app has its unique challenges, and the "soft skills" around people and process management become even more critical.
But these skills are important for efficiently running product at a SaaS org, too 👇
2. AI changes the revenue retention game
Newsflash: Salesforce goes all-in on consumption-based AI revenue 🤖
Two weeks ago, Salesforce turned the B2B software industry sideways - for customers, competitors, and their own business - when they announced the new pricing model for their agentic AI products.
The model is called "Flex Credits." Some people are loving it, and some are hating it. But the model's two biggest implications for SaaS are clear:
It's a consumption-based pricing model (not seats-based)
It provides a way for customers to easily shift their spend from seats to credits (called the "Flex Agreement")
The big picture 🖼️
Salesforce, a SaaS industry bellwether, is morphing from a subscription revenue business to a consumption revenue business.
The rise of AI SaaS products made this trend inevitable. As SaaS apps shift from simply enabling human workers toward actually working for them, it makes sense that the number of humans on the product won't be the best measure of value. Production and results will be.
Eyes wide open 👀
SaaS companies can and should embrace this shift toward consumption revenue. But they've got to do it carefully, because the subscription vs consumption trade-off has big impacts on the financials...especially revenue retention.
The top 5 financial metrics for ProdOps
At our last Office Hours session, the group discussed financial metrics for product leaders - which ones are most important, why they matter, and how product teams impact (and are impacted by) them on a daily basis:
ARR Growth 📈
Gross Retention 📌
Net Retention 💸
Gross Margin 📒
EBITDA 🏁
(BTW, our next Office Hours is on 6/5 - register here 🔗 if you want to talk about using AI in the development lifecycle)
These 5 metrics are how strategically valuable Product teams solve for the business outcome. See the breakdown 👇
Last edition we discussed the profitability metrics. Meanwhile, ARR is all about growth. Today we're digging into the retention metrics - aka profitable growth.
Let's ~try~ to define revenue retention
As standard as rev retention has become in the SaaS industry, it is still a non-GAAP metric. If you look at 20 annual reports from 20 SaaS companies, you'll find 19 slightly different definitions (the last one just doesn't report it...👎).
But here's the crux of it:
📌 Gross Revenue Retention (GRR)
The annual revenue you’re getting today from the customers you had 1 year ago, not accounting for any expansion revenue or new customers, but definitely accounting for any customers you lost. Min is zero, max is 100%.
💸 Net Revenue Retention (NRR)
Take GRR and now factor in expansion revenue as well, but still not new customers (remember: retention is a cohort analysis). Min is zero, max is (theoretically) unlimited.
To better illustrate, here's a hypothetical company with 3 imaginary customers, and what the GRR and NRR would look like in 2 different year-over-year scenarios: 👇
Mind the gap 💔
The difference between gross and net retention for Scenario 1 is just 7 points, but for Scenario 2 it's 18 points. Just from looking at this, we'd guess that Scenario 2 is the younger company, and it's still searching for product-market fit. Some customers really love it, but others really don't. Both customer and business outcomes are highly inconsistent. The more mature company in Scenario 1 may be less flashy in NRR, but it's more sustainable since it's backed up by a tight GRR.
So what's "good?"
For GRR, the conventional wisdom is simple: mid-80's to high-90's, depending on your product and your ICP. The low end is healthy if you sell to SMBs, who churn easier. The high end is expected if you sell "keep-the-lights-on" types of products to mega-corps.
For NRR...it's more complicated. As a baseline, 100% is "barely alive" - the minimum for a viable SaaS company. Your install base is self-sustaining, but all your growth has to come from landing new customers. And the bigger you get, the harder this is (plus it's more expensive, and more sales/marketing spend hurts margins). So "good" is as high above 100% as possible.
But what's realistic? In a lower-growth environment, 110% NRR is healthy. Just like how the Age of Efficiency changed growth and margin profiles, it also changed NRR - see the downtrend below:
OK, but what's AI got to do with this? 🤖
While AI agents are new, consumption-based software pricing is not. Infrastructure companies like Snowflake, MongoDB, and Datadog have been running this model for years. And they have a critical lesson for any subscription SaaS company making the shift:
Consumption pricing is leverage for your NRR.
It helps you on the way up...but hurts you on the way down.
Case in point: while the industry median saw about 10 points of NRR compression from peak to present, Snowflake saw over 50 (178% Jan '22 ➡️ 124% Apr '25).
As SaaS companies make the shift toward AI apps and consumption pricing, they'll have to carefully manage NRR. And product teams play a big role in that.
Retention isn't just CS' job
Revenue retention is about customer value. Customers leave when their users don’t get value out of your product, but they up their investment when the users do.
📌 Driving GRR means keeping the customer around, accomplished via...
Effective Onboarding: Shortening time-to-value with guided checklists, role-based tours, and auto-setup wizards. ProdOps Impact ☑️ Own the “aha" metric dashboard, monitor laggards daily, and trigger in-app nudges before CS ever opens a ticket.
Health Scoring: Combining usage breadth, depth, and sentiment into a single risk score that warns early if a customer is at-risk. ProdOps Impact ☑️ Pipe product data straight into CS tooling, surface red accounts in a weekly Product × CS huddle, and fast-track fixes inside the sprint.
Voice of the Customer: Embedding feedback widgets into the product and following up on everything you ship. ProdOps Impact ☑️ Route requests to a triage queue, auto-notify voters when status changes, and close the loop with a “Did this solve it?” micro-survey.
ICP Fit: Keeping poor-fit deals out of the pipeline and right-sizing terms for the deals you do land. ProdOps Impact ☑️ Feed churn post-mortems back into lead scoring, tighten discount guardrails, and give Sales a “red-flag” checklist at handoff.
💸 Driving NRR means getting the customer to want more, accomplished via...
Tier Upgrades: Figuring out which features represent a step-up in value, and thus warrant a step-up in spend. ProdOps Impact ✅ Meter usage of premium features, surface usage caps at 80%, and make the upgrade flow self-serve + instant.
Product Cross-Sell: Finding places where coordinating multiple use cases creates compounding value for the customer. ProdOps Impact ✅ Show contextual “You might also like” cards after a workflow is completed, and track attach-rate as a north-star expansion metric.
Seat Expansion: Converting users into champions and evangelists who want more people from their team/org on your product. ProdOps Impact ✅ Add lightweight invite prompts when weekly active usage peaks, reward champions with early-access perks, and measure viral coefficient alongside MAU.
Usage Expansion: A product manager’s dream, as this directly syncs usage/adoption metrics with revenue outcomes (the AI app / consumption pricing trend has the biggest effect here). ProdOps Impact ✅ Provide real-time usage dashboards, budget alerts, and forecasting tools so customers scale confidently rather than panic-cancelling
💥 The Takeaway: GRR limits your downside while NRR fuels your upside. But both start in the product.
Did a friend forward you this? Subscribe to Beyond The Click and get the latest on Pendo and Product Ops straight to your inbox every other Thursday.⚡
