The Platform Story: Part 4
Previously on…
Parts 2 and 3 were about goal frameworks and expectations — how to find and lean into a goal-oriented mindset, and through it, orient your organization around what it will take to build a platform.
This installment is about measurement, and it’s a little long, and a little dry, so there’s a tl;dr at the top. 🎉
Not covered here is the deeply personal, critically important, and entirely squirrelly topic of “why.” That’s going to need a separate post. And because it’s not really a “playbook” topic, it will fall off-cycle, and because I work in SaaS in 2019, it will bear heavy reference to the purpose-driving side of John Doerr’s OKR framework.
In other words, poke around the Measure What Matters site, watch the TED talk, and you can pretty much skip reading that post, whenever I end up writing it.
Anyway, back to the boring stuff —
Tl;dr:
Risk management in product development involves maximizing predictability with the majority of your resource time, and then making wild bets with a sliver of it.
Platform development is inherently riskier, but there are still tactics to measure impact directly using tools that (probably, and should) already exist in your organization:
- Usage: build product dashboards.
- Marketing impact: track your partnerships in one place and use campaign parameters for inbound links.
- Sales pipeline: get friendly with your sales ops, but don’t expect sellers to do your paperwork for you.
It’s a lot of legwork, but the earlier you get out ahead of it all, the better off you’ll be. Promise.
Details below.
Risk: the final expectation
No matter how well you set expectations, building anything carries intrinsic risks. You can do quite a bit to minimize risk of failure, but in the end, you can control only so much, and you’ll hit only so many of your strategic bets. The good news is, we-who-love-Platform can learn a bit from more traditional product development here, starting by examining how each discipline approaches and manages risk.
Risk in product development
Heads up: slight oversimplifications incoming —
In developing a product that generates direct usage metrics and revenue, you’ll generally want to balance predictability and risk. You have an audience that you’re trying to grow, expand, or retain, and everything on your roadmap aims to push some key measurement that’s tied to one of those outcomes. Your methods, frameworks, and overall approach may vary, but chances are you have a desired impact, and at least one dashboard that measures it.
In this world, the bulk of your available resource bandwidth is committed to the pursuit of predictability — shipping velocity, measurement, growth, revenue. And then a small bit goes toward wild exploration that energizes your organization, excites customers, and forces you to reevaluate the assumptions underlying your primary roadmap. This is the fabled “10% time” for things like “hack days,” and “R&D.”
This is obviously more easily said than done, but the theory isn’t rocket science: have ideas, plan them out, measure their impact, and make sure to leave time for experiments.
Fundamentally, you’re minimizing risk by maximizing measurable predictability; and then making bets with 5–10% of your resource time.
Risk in platform development
Platform is similar, but because your goals include getting other people to build things you didn’t think of, your landscape is inherently less predictable. Yes, you can make tools-as-features and measure their usage and impact relative to your core product, but your impact on the bottom line is less direct — in a way, platform creates the software equivalent of soft political power:
- Great tooling enables your sales org to smoothly navigate objections (“well that’s not on our roadmap, but with our API…”)
- A broad partnership story enables your marketing organization to leverage channels outside your company.
- A strong platform offering enables a level of bespoke customization that helps your core product team stay on mission, and minimizes strategic thrash.
- An ambitious platform roadmap enables your customers to imagine your company in a more personal light.
But these things are all a bit squishy, and getting good numbers relies on many departments coordinating on where metrics should live, and what they should mean. And while you should absolutely strive for that coordination, know that it’s likely to materialize less quickly than you’d like it to, and in the meantime, you’ll have to make decisions anyway.
Platform inherently skews toward riskier moves with less predictable outcomes — embrace a mindset of experimentation, and accept that nearly every decision is a bet.
Measurement
It’s not all wild west out there! With platform development, there are a number of things you can, and should, measure directly. In addition to helping set your next round of goals, metrics will serve a couple different purposes for you, so it’s important to get them in place early, and refer to them often.
Metrics can help with:
- Measuring impact: are you doing the right things? Is your work effective?
- Internal storytelling: your organization might not always understand what Platform is, what your goals are, or whether you’re successful — numbers help.
- Deciding where to go next: this is where cross-department coordination can really help. In a nutshell: de-risk your bets by using numbers to help spot your safest bets (or highest upsides).
Keep in mind: you won’t be able to do this alone. Remember that whole conversation about building consensus and setting expectations in your org?
It’s time to make good on all that. Tracking everything yourself may seem pretty doable at first, but you’ll inevitably end up in a cluster of too many surface areas in which you don’t have enough expertise or control, and it’ll turn into a tail-chasing exercise. So ask for help.
Anyway. While every situation will have its nuances, there are three big buckets that will apply to every organization: usage, marketing impact, and sales pipeline.
Usage
Raw usage is your easiest starting point. If some % of your user base are uploading content through your API, and your gross numbers all go up (i.e. it’s unlikely you’re cannibalizing), that’s probably a good sign.
At Frame.io, we’ve approached this in a number of different, imperfect ways:
- Google Analytics and separate event reporting on our developer portal
- Data pipeline usage of our API attributed to external sources
- Product dashboards for resource creation via API and OAuth tokens
I’d wager that for the age of our product, our measurement instrumentation is probably better than most — and it still kind of sucks. The pipeline usage is neat, but noisy, everything else is a bit gappy, and all of it’s tough to collate.
The big lesson I’ve come away with here is twofold:
- Leverage all your existing systems for measurement. More data is better, and speaking existing languages in your org is usually the right call.
- Pick one, or at most two things to focus on. It’s great to have lots of data, but trying to move and track every needle at once is likely to become distracting and frustrating.
My advice: for usage, build one or two product dashboards. It will force you to tell a story that the rest of your organization understands how to hear.
Marketing leads
This was a blind spot for me as we launched, and a year in, I’m still playing catch up. It’s also a pretty easy thing to figure out and get in place, which honestly makes me all the more annoyed with myself.
Here’s an ice-cold take: if you launch a major strategic partnership, and you earn loads of media impressions off the back of it, that’s almost definitely a good sign.
And that’s about where I stopped thinking and moved on.
You need to be able to measure this stuff somehow — otherwise it’s just anecdotal, and while it might improve your organization’s overall impression of your work, it’s not much of a bargaining chip when you really need to ask for help or make a case for yourself, or your team.
This is a two-part puzzle:
- Track your partners, and whatever they build, in one place. It doesn’t matter where, so long as you’re consistent. After about a dozen false starts (docs, spreadsheets, sync meetings, CRMs), I ended up settling on Airtable — I’ve found its feature mix to be just about right for my needs and scale.
- Your responsibility is at the top of the funnel. Marketing has (or should have) its own mechanism of tracking leads → sales → closed dollar. Trying to solve that from the platform side of the fence is paralyzing. Just don’t. For the purposes of this exercise, your job is just helping expose whether a lead came from partner media. So talk to marketing up front, figure out a strategy for campaign attribution, and follow it.
It’s honestly that simple. And yes, it’s a bit of legwork. But it’s critical, rewarding, and a hell of a lot easier to maintain in flight than to build from scratch after a year of flopping around without proper marketing attribution,
Max, you idiot. 🙄
Also, don’t forget to be a good citizen — ask your partners for their campaign trackers. After all, there’s a good chance they screwed this up, too.
My advice, in case it isn’t obvious: track everything in one place, and solve for campaign-level attribution for partner media nice and early.
Sales pipeline
Ah, pipeline attribution. This honestly deserves its own post — and may eventually get one — so I’ll be brief here.
The fundamental driver for this is that the closer you are to revenue, the more obvious your influence on the business, and the more stake you can build. And that’s great. But you shouldn’t expect anyone to go out of their way to help you drive that attribution, so make sure you set aside extra time for:
- Training your sales team about what you do, and how you can help on deals.
- Socializing any changes you need to drive into your company’s sales CRM of choice.
- Inevitably doing a lot of your own paper-trailing and reporting.
The upside is that you might get to spend time with your sales ops and enablement teams, who are — if your experience is anything like mine’s been at the last few places I’ve worked — fantastic coworkers.
In our case, we’ve implemented some fairly nuanced fields and custom objects so we can collate deal-stage-based revenue potential to help prioritize new partnerships. You might not need that. Honestly, we might not need that.
My advice: at the very least, you’ll need a field that tracks whether or not your platform tooling is critical to a deal, and to what degree. That data will help you drive attribution, prioritize your time across deals, and ultimately, forecast headcount.
See, aren’t attribution and measurement fun?
Don’t answer that. Thanks for reading, and see you next time.
Previously:
Coming up:
(By the way, this list is entirely subject to change).
- Part 5: Choices, mistakes, and lessons.
- Part 6: More choices, more lessons.
- Part 7: The partner pivot.
- Part 8: Keep your head in the game.
- Part 9: TBD…
