This is the third in a series of posts on Growth Marketing on an enterprise scale, based on my experience leading several growth projects for Microsoft Azure and Amazon Web Services. In this post, we traverse the steps 3 and 4 of the growth pyramid I defined in the part 2 of this series, as a tool to solve the main challenges every large organization faces to set up growth / growth marketing teams up for success.

In the previous post, we discussed ways to find the jobs your customers will hire your product for, and define a North Star Metric (NSM) and KPIs that will help you measure the efficacy of your growth programs. In this post, we’ll discuss the baby steps growth teams need to take in an enterprise setup to evolve into a full-fledged, sustainable function that contribute meaningfully to growth for years to come.

“Run when you can, walk if you have to, crawl if you must; just never give up.” — Dean Karnazes

I haven’t always been the biggest fan of the “crawl, walk, run” approach, believing it to be an excuse that teams sometime make for the lack of velocity. My career as a growth PM and marketer has made me value this approach a lot more than I could ever imagine. In an organization where growth is not a well-defined function, you’re unlikely to have the luxury of a growth team with dedicated product, data science, and marketing resources at their disposal. Success for a growth PM will hinge on the ability to create a “virtual growth team”, which at times will pull people away from their day to day fires or their priorities. The only way this model will sustain is if it’s bootstrapped with some early success.

That brings us to the next steps of the enterprise growth pyramid

Step 3: Choose channels and partner teams, drive alignment with partner teams

I’ll use the example from my previous post — While I was growth PM for Azure, we knew that customers commonly find and adopt Azure first through the Visual Studio IDE, the environment where they write the code for their application. Azure and IDE are distinct teams at Microsoft, reporting up to different VPs. Knowing that we’ll need to rely on the Visual Studio team to drive Azure adoption, the first step was to pick KPIs that would contribute not only to our NSM but also to the IDE team’s priorities. We talked about that in the previous post. Choosing the right KPIs is, however, not sufficient. We needed to find a mechanism to make the IDE team own these KPIs as much as we did. So we asked our leadership to help us report on these KPIs in the IDE team’s monthly reviews with their own leadership. By doing so, we would signal our intent of helping the IDE team’s priorities, while at the same time providing visibility to the team on how these KPIs would affect our growth program and raise red flags/alleviate blockers whenever they occur.

The second step was to deploy our limited resources on the most effective channel for growth. The IDE team lacked a robust notification engine and

we didn’t want to rely on an out-of-context channel, such as email, to nudge the user to the next step when they acted on a “trigger” (identified in the trigger-action-reward steps from the previous post). So, we spent the first few sprints just testing and shipping a notification engine that had been developed by the team in the past, but later shelved. Meanwhile, to ensure that we’re utilizing this time well, we created high fidelity design mocks of the experience we had envision for the test. Using these, we quickly got user feedback from usertesting.com, which helped us make some early, low cost revs to the planned test. If we had launched a growth pilot with existing channels that may not have been the most effective at nudging users along, we may have jumped to the wrong conclusions about the program in general and perhaps even lost support of our partner teams or the leadership, which were both obviously skeptical of this new approach.

To quickly summarize — driving alignment with partner teams by starting from the NSM and using the KPIs that they would care about too and choosing the best channel for the growth experiment are critical steps in the context of risk-averse large organizations with a fledgling ‘growth’ function.

Step 4: Pick a low-hanging fruit and complete a growth experiment

While this step may be obvious in theory, the execution in practice can be rushed. Most teams in large organizations don’t spend sufficient time in building a solid backlog of growth ideas before they can take an educated decision on which ones qualify to be the lowest hanging fruits. Instead, they jump to the first few hypotheses that come up in a meeting and start to pursue them. If the chosen experiment takes a lot of resources and doesn’t show promising results, the excitement for growth projects can quickly fizzle out.

So how can you and your team build a backlog quickly without going into a lengthy design sprint? Get into a room with your team and start putting up some post-its with ideas. culminating in a straw poll to prioritize them. The secret sauce here is to let creativity flow, and say “Yes, and..” more than “No, but…” — a lesson that was taught to me by an Improv group in LA during business school, and one that I constantly use in brainstorming meetings till date. You could also do it in a more structure way with a 90 minute to 2 hour long design thinking meeting, as illustrated in a blog post by DesignMap Inc.

Sometimes low-hanging fruits are hiding in plain sight — in excel sheets or dashboards.

If you’re trying to grow a specific part of the product funnel, try to segment it in different ways and you may find your most promising growth idea. Sometimes low-hanging fruits are hiding in plain sight — in excel sheets or dashboards, and you don’t even need a design thinking meeting. In my current role growing AWS’s data warehousing service called Amazon Redshift, I was initially tracking the conversion rate from trial to paid usage in the funnel (we have a 2 month free trial). With a healthy 40% conversion rate, I had pre-maturely moved on to exploring other areas in the fall of 2018. During the holidays, though, I got a chance to come back to this part of the funnel and take a few different cuts — including data for the past 3 years. I found that the users who had any paid usage during their free trial converted to a paying user at a rate of 70%, while those that did not pay converted only at 25%. The latter segment represented the majority of the users trying the service, and I beat myself up for not noticing that earlier. With a quick back of the envelope math, I concluded that by just increasing this conversion rate by 10%, I could create a $5M lift in the monthly run rate for the service. No prizes for guessing what area of the funnel I focused on over the next few months.

More on step 5 to follow in the next post in the series. To conclude, we laid down a 5-step plan to start setting up a growth product/marketing team and discussed two of those at length — choosing channels and partner teams, and picking low hanging 🥝. Super eager to hear your stories on how you think of growth in this context and how you picked some of your first growth experiments👇 .

Ayush Jain is a problem solver first, and a Marketer or Product manager second. By starting from the user and understanding the most effective lever to solve a problem, he works across product and market levers on growth projects that create value for users and move them along in their journey with the product. He currently works on growth for Amazon Web Services, and previously worked as a growth PM for Microsoft Azure.

. Super eager to hear your stories on how you think of growth in this context and what are some of your NSMs 👇 .

. Super eager to hear your stories on how you think of growth in this context and what are some of your NSMs 👇 .

I’ll start with Sean Ellis’s framework for growth for young companies and try to evolve it for the enterprise based on my experience

You can see how the process looks eerily similar, but adapts to the maturity of the organizational structure. The areas with stark differences are the amount of alignment you need with partner teams to build up a culture of growth and the process adaptations you’ll have to painstakingly vouch for against the inertia of processes that have hardened like concrete over several years. Once you’re able to break through these barriers, the enterprise scale would work in magnificient ways for you. Just think about the speed of your growth experiments with a product that already has millions of tire-kicker users and millions in revenue run rate (at Microsoft, this happens frequently very shortly after a new product is launched because of the halo effect of other related products). Let’s take this step by step:

Step 1: Research your customers

At the risk of sounding like a broken record, I’d still always advise we begin with the most basic step. While you may already be doing a bunch of customer development, interviews, focus groups, etc. to uncover product ideas — this research may not always give you great pointers on how to start thinking of growth. So the obvious question is — how do I adapt this research for growth?

As a product thinker, I constantly go back to two fundamental approaches of behavioral economics that lay the foundation of my growth projects —

1. I use Bob Moesta’s Jobs to be done (JTBD) to segment users based on the real behavioral trigger points of why users even give our product a shot, or press the 💲 button. This helps me get a sense of how growth happens at the top of the funnel and what levers can be used to spur more growth there, depending on the target audience. This may be a “meh” point for a lot of you valley readers who’re hearing about JTBD more and more, but you’d be surprised at how few product/marketing teams think this way in an enterprise setup. If you’re new to this, a simple example is the following — what would you do to grow Instagram’s user base? If you go and interview users who joined Instagram you might get simpler answers, such as a high % of responses telling you they joined because a friend invited them or got them hooked onto Instagram. And you might think that driving more referrals from the platform is an effective strategy to get more users. However, if you dig deeper, you may find that a user went through several stages that ultimately snowballed into a desire to check out Instagram when they received an invite. An invite sent before this “magic moment” may have not successfully done the job. The user may have seen 3–4 posts on Facebook from Instagram because a lot of their friends are already using both platforms, they may have come across news articles about celebrities who posted on Instagram, etc. This user has reached a different level of motivation than a user who have only 1–2 of their friends on Instagram and otherwise may not have come across Instagram “passively”. I’d target the former users with “referral” experiments.
   How does this translate to a B2B product? An example is what I found when I was leading growth for the then fledgling Visual Studio Team Services (an Azure Devops service for developers). One of our big challenges was gaining share from Atlassian’s set of Devops services which were immensely popular around the time Atlassian IPO’d. We were spending a lot of money on direct response marketing (DRM), sponsorships, PR/AR — but were struggling to hit our growth targets. It was only in 2016 that we started to interview new Atlassian users about their journey to finding, trying and ultimately buying an Atlassian product (in contrast to only interviewing habitual users of our own or competitor products). We found that a large % of them had been following Atlassian’s organic content on topics such as Git, Agile planning methodologies, Devops practices, etc. They were never searching for Atlassian but just trying to keep up to date with the latest in industry practices on the jobs they were trying to get done, which Atlassian had smartly captured in content that was consistently ranking high on Google. And because they were following this content on Atlassian, they naturally got introduced to the Atlassian products. This ultimately led to the dominoes falling and them trying these tools out and advocating for them with their teams. Based on this realization and the discovery of the most common jobs that fed this funnel, we completely re-thought our awareness and DRM strategy, leading us to publish Devops content that’s unrelated to our product per se, but targets top of funnel keywords. Our monthly user signups grew ~50% since this content started ranking high in SERPs. Enterprise software doesn’t just have a long sales cycle as is the common perception — it also has a long awareness and trial cycle. It’s more important for a B2B than for consumer product to drive users down their consideration funnel from the least aware to the most aware stage (My go-to strategy here is Eugene Schwart’s five levels of awareness). A B2B growth strategy really needs to factor that in.
2. I use Nir Eyal’s Trigger-Action-Reward-Investment framework to structure my research for growth projects that target engagement and retention. If JTBD is a good way to uncover levers that’ll create the “big hire” (as Bob Moesta likes to call it), the Nir Eyal framework opens your eyes to each “little hire” i.e. what triggers your users to return and use your product. I also like it for its ability to train you as a growth marketer to couple your insights from the customer “triggers” with clear “actions” that you might want to leverage in your growth strategy. So how does this land in practice? I’ll use an example from my time on VSTS again. Even though 50K+ users signed up for the service each month, churn was a big problem for us initially. We’d validated our early hypotheses with quick call downs with customers who’d churned quickly after signing up, and one of them was the obvious — an empty state experience that didn’t guide users to the magic moments. We tried a bunch of experiments to create a guided experience for new users which only resulted in a slight increase in first week retention and didn’t quite make the dent we expected. Further digging into the usage data, we found that the churn was more pronounced for lone users than for users who also had team members. While this may seem obvious in hindsight, we weren’t sure if a team begets engagement or engagement begets creation of a team. This is when we shifted our interviews to focus on uncovering the trigger specifically for users part of a team. We found that the notifications being generated by a team-mates’ code commits and work tracking changes caused a user to login to their project again, and therefore being part of a team caused more engagement 💡. While we hadn’t designed the experience to create an external trigger for a user, it had the happy side effect of “triggering” the user’s interest in what their team-mates are doing, leading them to “act” on it by checking on the activity that just happened and feeling “rewarded” on seeing that their project is making progress. Knowing that this T-A-R flywheeel had already been established somewhere in the product — we started more proactively driving users towards that flywheel by nudging them to invite their friends earlier in the lifecycle, rather than thinking of creating new flywheels through other levers. This is another interesting nuance often noticed in B2B software more than consumer software — harnessing the power of teams to create user engagement cannot be underestimated.

Note that I swapped out Product/Market fit (PMF) for customer research here. This is due to the assumption that big companies start investing in marketing for a product only when a certain adoption threshold has crossed (direct value)or synergies with other products (strategic value) realized. Without these two happening, rarely do product teams start scaling and the tipping point for starting growth teams isn’t quite reached. PMF is perhaps more important to call out for startups to prevent them from falling into the temptation of investing in growth way too early.

Step 2: Defining and socializing the North Star Metric and KPIs

Once you have a good idea of your users’ intrinsic and extrinsic triggers to “hire” your product, you can start to form an idea of the least common outcome across these triggers. This will be the foundation of a north star metric (NSM) that will guide your growth projects.

The NSM for your growth projects may be slightly different from your product or company’s NSM. For instance, if you’re part of Airbnb whose NSM is nights booked but you may be running a growth team that’s laser focused on growing the supply side (bringing more housing inventory to the platform). The least common outcome for you, then, could simply be # of new units added per city per month. It’s useful to have a more focused NSM than the one your company/business tracks for a few reasons:

(a) It creates more focus for your team. There’s simply too many ways to influence the “nights booked” NSM that might cause randomization within a new growth team. Focusing on one aspect aligns all arrows.

(b) It gives you a better statistic for quantifying lift and p-values in your experiments because it’ll relate more to your actual growth levers. A well functioning growth team on the supply side might not be able to influence the overall Airbnb NSM if the demand side is weak, but should not invalidate their promising supply-side experiments by measuring a statistic that has too many confounding variables (the overall NSM).

Another note is to make sure you avoid a common enterprise pitfall — choosing revenue 💵as your NSM. This is because growing revenue doesn’t always indicate growing usage, especially in companies with a large enterprise sales force or partners. Thinking about the early days of Microsoft Azure, we commonly called out the 3-digit increase in cloud revenue to position ourselves as a strong contended in the new computing paradigm. While this wasn’t a false claim, it create a smoke and mirrors effect on what was really going on — our sellers were leveraging existing relationships and deals for Microsoft Office, Windows and other products to add a few thousand $s of Microsoft Azure commitments in the company’s Enterprise Agreement when it came up for renewal. This was a pretty smart move as well, and could have been a really good incentive for customers to try the new cloud platform. However, this tactic wasn’t successful at driving actual usage of the committed 💲. It also created a slippery slope for the marketing teams — who were happy to create more offers and programs to drive these deals, but weren’t focused on driving usage. In 2017, Microsoft’s sales incentives were finally re-structured to focus on driving consumption vs driving sales. Our NSM is now consumed 💲, which creates better incentives for all teams 😁.

Again, in practice, how do we think of an NSM for our growth teams? When I started to run growth experiments for our developer subscriptions (also known as MSDN), only a small % of paying users were using Azure and we were measuring this statistic directly. However, through step 1 we realized that the most common triggers for the jobs subscribers hired Azure for happened during their use of the Visual Studio IDE (the hero product in the subscription). They may use Azure to deploy their application, or to create a SQL database to use in their app, or simply to spin up a VM where they can test their app. All of these jobs were triggered while they wrote code in the IDE, and led them to create an Azure account and become engaged on the new platform. We therefore changed the NSM to focus on the the least common outcome — % of engaged IDE users who are engaged Azure users.

The next important step alongside the choice of NSM is to have KPIs for your team that’ll actually help in collaborating with partner teams. KPIs should be the most important metrics that influence the NSM and can be strategic levers to drive alignment with partner teams. So this will be super important to get through step 3. For us in the above example with developer subscriptions, we knew that we had to take users through a few important steps before they’d hit the eventual north star goal. For instance, we had to ensure a healthy rate of paying subscribers were engaged on the IDE to satisfy the fundamental premise of our strategy. We also had to also track the health of incentives we were putting in place (such as a $150 monthly credit to incentivize trials of Azure). And finally, we need KPIs specifically to spur collaboration with the product team. So we also wanted to take accountability for 1–2 KPIs that might only indirectly accrue to our NSM, but provide credibility and a ‘place on the table’.

The relationship between KPIs and the NSM can also be thought in terms of a growth equation for your product/business such as the one (A growth equation is also a great validator for whether or not your product has the potential to grow scalably and if not, what key variables are missing)

You can swap the right hand side of this equation with a NSM for Amazon like “# of orders/per month)

More on steps 3–5 to follow in the next post in the series. To conclude, we laid down a 5-step plan to start setting up a growth product/marketing team and discussed two of those at length — customer research and setting up a North Star Metric and KPIs. Super eager to hear your stories on how you think of growth in this context and what are some of your NSMs 👇 .

Here are some useful resources:

1. Intercom’s Jobs to be done ebook is a great primer and a quick read
2. Sean Ellis takes an in-depth look at the north star metric
3. Andy Johns’ indespensable growth frameworks is a great guide on how to think of talent for a growth team.

I’ll leave you with a diagram by Sean Ellis that illustrates the NSM and KPIs for LinkedIn’s growth team

Ayush Jain is a problem solver first, and a Marketer or Product manager second. By starting from the user and understanding the most effective lever to solve a problem, he works across product and market levers on growth projects that create value for users and move them along in their journey with the product. He currently works on growth for Microsoft Azure.

Growth marketing today is a bit like Product Management used to be 10 years ago — a rapidly growing and often misunderstood function that’s highly variable across organizations. It’s also a unique skill set that’s in rather short supply — equal parts Data Science, Marketing and Product Management. Yet, we’ve come a long way since the time Facebook pioneered the practice to grow its ~50M MAUs to 1B+. There’s some amazing blogs, podcasts and even a dedicated conference about the practice (some of which are listed at the end of this post). LinkedIn pegs the no. of Growth Marketing jobs at 7.5K+ in the United States, vs. 25K+ for Product Marketing and 115K+ Product Management jobs. That’s pretty amazing growth for a function that’s relatively new in terms of being an official job title.

However, the more you dig into it, the more you’d observe the examples, best practices and thought leadership to be focused on companies or platforms with small to medium scale — from Stage B to just over post-IPO scale. (Note: Growth marketing is most effective when a product finds product-market fit, adoption reaches a critical mass and churn has stabilized to manageable levels. So you don’t see a lot of examples at very early stages in the maturation cycle). So what about enterprise companies such as the FAANGs, Microsoft and others? As a growth marketer at Microsoft, I constantly grapple with problems similar in nature to the ones growth marketers in the valley are trying to solve. However, I constantly learn the hard way that the same solutions are not effective at enterprise scale, irrespective of how much product innovation your company may be driving.

I’ve learnt that there are three main challenges that make Growth Marketing less likely to be effective in an enterprise. Carefully auditing and removing bottlenecks in these areas in your organization could position your growth team for success much better. Some of these may seem obvious, but their impact on growth activities is more magnified than on day-to-day product and marketing activities.

Functional silos

Product and marketing teams in a software company spend most of their time doing two primary activities, respectively —(a) building a product people love and (b) growing adoption and/or sales efficiently and with the ability to scale. In companies where product and marketing are part of separate units (such is the case with Microsoft, Google, Amazon, etc.), efforts across these two key activities often become parallel workstreams, instead of being two levers of a single workstream. This causes each function to run growth experiments with the levers available only to them, even if they may not be the most effective levers at driving a certain outcome. For instance, marketing teams at Microsoft routinely set up emails for user onboarding and engagement. But often these emails are not accompanies by a solid onboarding experience in the product, leading to a high churn rate (emails are out of context and not always the best onboarding mechanism if delivered standalone). On the other hand, engineering teams that track churn as a KPI routinely try to do in-product optimizations to improve churn, even though the most effective tactic here could be a proactive email sent to a user who may never return to the product again, but may still read the email.

Obviously, placing a growth team in a specific functional unit would not be effective.

How do you create a growth team that’s more cross-functional?

Data silos

The reason why growth marketing at Facebook is so effective is that they have a consistent user id (a 128-bit UUID) across all products such as news feed, photos, etc. Designing your data schema around this single unit of identity is often a far-fetched dream for enterprises like Microsoft, that are moving from desktop software to cloud software, devices, etc. Further, if the company happens to be in the B2B space, a user is not always a well defined entity — often several separate users may come to visit your website, a different user may decide to purchase your software, yet another user may actually buy and administer it, and a completely different set of users might end up using it. Various commerce engines and marketplaces, GDPR protections and a plethora of other issues make the story far more complex here than for a SaaS company.

Who is your end user? Getting to an identifiable user is incredibly hard! Even if you do — all the best trying to work with managing opt-in/opt-out.

Inevitably, a large company will drive some standardization of its data infrastructure across its product, marketing and sales teams with the best intentions. For instance, a lot of large companies are choosing to dump all of their product telemetry, commerce and user PII data into a data lake. Now, what if a team wanted to use a combination of this data to create an email campaign? This is incredibly hard with existing tools for running email campaigns (Intercom, MailChimp, Exact Target and others). Most email tools like data to be stored in structured SQL databases. Marketers also do not have the skill set to query data lakes directly to build their campaigns. Finally, if you do use a third party tool, you’ll have to find a way to sync user’s opt-out of promotional emails back with your company’s permissions master database. Phew!

How do you work with this clunky infrastructure? You can kiss agility goodbye!

Inability to use the right tools for the job

Data silos is not the only issue that comes in the way of using the right tool for the right job. Big enterprises are more wary of user data when it comes to using it to offer personalized or tailored experiences — which is a key tactic for growth. This wariness is even more pronounced now under the GDPR regulations. For instance, Microsoft’s Privacy and Cookies terms commit to not using cookies for re-targeting users with display ads on ad networks. This hurts any team that’s looking to do demand generation or drive engagement for existing users, since re-targeting could be a very effective technique for both scenarios.

Data governance and protection issues also often act as a barrier to teams adopting best of breed tools. Just buying a subscription for Intercom was a 1.5 year journey for us, because we had to cross several privacy, accessibility and legal validation checkpoints for scenarios in which Intercom was going to have access to our user data and telemetry. A growth team would also need tools such as Google Analytics, Optimizely for A/B testing, Usabilla for user feedback, Segment to combine data from disparate sources and others across the user funnel. Because all of these tools need data to be stored outside the company’s firewalls, using them is rarely a trivial decision point.

How do you get around the inability to use promising tools and still iterate on growth experiments?

The final picture

What you ultimately end up is a world where several teams are reaching out to the same user through their own channels, but unaware of the other channels that may be trying to drive to the same outcome. With luck, you could have a constructive effect of these efforts. But most often it results in the user getting spammed with disparate messages. If we try to depict this in a more visual way, you end up with something like below.

My intention with this series of posts is to highlight how growth marketing in enterprise companies needs some more key ingredients than growth marketing in newer, smaller companies. However, by no means does that mean that growth marketing is not your cup of tea if you’re an enterprise and B2B marketer. With creative organizational alignment and growth tactics that borrow from industry leanings but adapt them to your organization’s unique challenges, you can accomplish a solid ROI from your growth efforts.

Meanwhile, I’d love to hear your growth challenges. How do you think about growth in your org and what prevents you from being successful? Have you learnt creative tricks to get around these roadblocks? Please leave a comment!

If you found this post useful, please hit 👏.

If you’re new to growth marketing, here are some of my favorite resources:

1. Anu Hariharan’s Growth Guide on YCombinator blog
2. Andy John’s frameworks for growth
3. Intercom’s podcasts and blogs on Growth
4. Sujan Patel’s blogs

Ayush Jain is a problem solver first, and a Marketer or Product manager second. By starting from the user and understanding the most effective lever to solve a problem, he works across product and market levers on growth projects that create value for users and move them along in their journey with the product. He currently works on growth for Microsoft Azure.