How RevOps changes in PLG

Inspired by this post by Sam Crowell Richard , I wrote this article with some of the best minds I know.? Most of them are quoted below but I deeply appreciate everyone who contributed.??

Drop comments below or DM me for the source-doc where you can comment in-line.? I’ll keep updating it as an ongoing resource.?

1) Data moves beyond traditional SaaS systems and your data foundation is critical for success

In a PLG world, you are no longer just dealing with account / contact / opportunity data that is well structured in traditional SaaS systems (CRM, MAP and Support are the traditional ones).?

In a PLG world, you need to connect the following data together across all your systems

• Sales data (salesforce accounts, contacts, activities)
• Marketing data (engagement, activities)
• Support data (tickets)
• Financial data (subscriptions, customer status)
• Application data (users, accounts and usage data)

You can’t leverage product data at scale with legacy CRM (Salesforce + Hubspot) and Marketing Automation (Hubspot and Marketo). ? Their architecture doesn’t support it beyond a few select product data fields. ? Companies have hacked workarounds for the core SaaS stack including:

• Moving product-triggered emails off Marketo (Iterable and Inflection)
• CSM which is driven by product data (i.e. health score) moved off a while ago to tools like (Catalyst and Gainsight)

“Legacy systems were meant to handle columns of data, not streams of data.? We see large PLG companies where the volume of data (events, contacts, API, activities, etc is so great you can't even put it in Marketo even if they wanted.”? Dave Rigotti @ Inflection

“The legacy platforms’ ability to handle event/usage data is definitely a challenge but ultimately I think the bigger issue here is reps really can't synthesize a stream of events or product metrics into action. The role of RevOps in PLG I believe is taking the raw unstructured data and transforming it into actions/playbooks for reps.”? Nikko Georgantonis @ Hightough??

But these hacks are just patches.? To be successful with your PLG motion, you are going to need a data stack , not just a GTM Stack

• You will need to have a data warehouse
• You will need to have data pipelines into the warehouse
• You will need to be able to model data
• You will need to be able to pipe that modeled data into key systems to not just inform but also trigger actions (see more below)??

“With PLG, you are flipping the model on its head. Instead of sales → product it is product → sales.? At Atlassian, we had millions of user records in our prod database; a very small percentage of these were in Salesforce. Existing CRMs aren't designed for that amount of volume of records or event data.”? Patrick Thompson , CEO of Clarify

Implementing a data stack sounds simple but it isn’t.? Data doesn’t map well between unstructured systems (i.e. product).? Here are a few common problems we’ve seen:

• Rarely will there be a 1:1 mapping between an account in the production system and in your CRM. ? You may have to build a junction object…?
• Transitioning a freemium user into a company account with multiple users creates complex data mappings
• You will likely have many users in your product that aren’t contacts in your GTM systems

“You have to be purposeful on the data you bring and how you pick up from an event stream and where aggregation happens and the frequency of updates.”? Mollie Bodensteiner

You need to partner with data / analytics / product teams to get access to and understand data.??

• Very rarely will RevOps own product data so cross-functional partnerships are critical
• Product data is much more detailed with tons of fields, often without clear definitions (without asking the engineer who built the feature).? Understanding the right field to leverage is critical to effective triggers.
• RevOps won’t have this deep product knowledge so you will need to partner with product
• You will also likely want a data capability on your RevOps team so you can effectively partner with data teams across the company and visual data in BI.?

Finally, we believe that a new wave of GTM tools, built on top of the modern data stack is coming !? You need to get ready for it!

“You should invest in systems that connect with the modern data stack to execute faster and reduce overhead (ex. Not have engineers dedicated to keeping Marketo <> Snowflake data pipeline alive)”? Peter Kirk , Head of RevOps and Professional Services at MadKudu

2) Triggering emails from data is a new capability RevOps needs to build

• You are no longer replying just on reps to send emails. With PLG you have product data to trigger emails.? Some examples of communications you can trigger include:
• Increasingly you can trigger and personalize emails easily… There is a world soon where everything but the call and reply (when a prospective customer replies) will be automated (separate article here - click to request access)
• If you want to create a seamless experience for your customers and prospects, you can’t have your product emails and your marketing emails separated (this is what Marketo / Iterable push today).? Instead you want them all in one system (the vision of Inflection)

“it's imperative that PLG companies move from time-based to trigger-based communications.? This is this the single biggest high ROI activity for free to paid and activation improvements.” Aaron Bird , CEO at Inflection

3) Scoring is no longer a nice-to-have.? And no longer just a configuration in marketing automation tools

• Being able to prioritize the right accounts and trigger actions to them is critical
• There is more to scoring than just a regression model in a data warehouse:
• You will also need to think about which teams you are routing leads to as there is a different motion for sales-assist and enterprise.

“We've split motions into "high-velocity" and "higher-touch" because we need to make sure that sales is spending the right amount of time on the right deals (price points can be low in PLG, so you actually want to try and predict the ACV of different types of customers). This can be tricky to get right and to path customers correctly and make sure sales isn't spending too much effort on low price point engagements.” ? LIndsay Rothlisberger , Head of RevOps @ Zapier

4) World class CPQ with self-service is foundational to in-product upgrades

• With PLG your pricing model is typically much more complicated
• You CPQ is no longer just be something reps access to create quotes but it will be critical to integrate with your product to allow customers to buy / upgrade themselves (you’ll typically find you only touch a small portion of your customers with reps in a PLG model).?
• Self-serve upgrades can drive a significant percentage of ARR and quota relief.? Note that sales incentives will also need to change to avoid competition between sales-assisted and self-serve motions.

“Creating an upsell process from self-service > Sales led is also not straight forward from a reporting/systems standpoint. If you don’t get the CPQ right, this could look like a churn + new business versus an upsell.”? ?Nikko Georgantonis @ Hightough??

“One of the biggest changes is that marketing ops 1) starts to deal with post purchase not just net new 2) as result needs to be involved in product usage / events ? Dave Rigotti @ Inflection

5) You can’t run your sales / CS teams without a tech-touch model

• With PLG you will touch only a small portion of your customers and prospects.? You need to automate as much of this as possible (both for CS tech touch and rep-assisted sales) See more thoughts in this other article

“We're automating a large portion of sales outreach because volumes are too high to warrant simply "handing leads off" to sales.? This requires an entirely new skill set where RevOps is using data to identifying opportunities in the funnel and executing on actually driving conversion.”? LIndsay Rothlisberger , Head of RevOps @ Zapier

• The roles of sales, success, support and SDRs will change and blur

“We're seeing SDRs book 80+ meetings a month with well-designed prospecting plays (i.e. consolidation of accounts, account nearing paywall, expansion across teams, cross-sell of products based on activity) when the average used to be 10-20 per SDR. This comes from having clear plays to run and full/semi-automation of those plays.”? Sam Levan , CEO of MadKudu

• Your segmentation and routing will also change in a PLG model.??

“While you still will want to map out your target accounts and assign them across your reps, there are going to be a lot of companies you didn't expect. And your typical "lead routing" isn't quite sophisticated enough because that assumes you have a discrete demo form fill that came inbound, whereas you really want to encourage allowing reps to outbound the "warm" accounts (but also not force them to hit the "warm" accounts with an SLA).” ? Tido , CEO at Koala

6)? Attribution is now 3-D chess

• With enterprise deals, multi-touch attribution has always been important as you want to understand how all channels influence a transaction
• Done well, SDRs, demandgen, & AEs complement your efforts and drive prospects to sign up for the product (and get the appropriate “credit” for doing so)
• In PLG you also need to determine what counts as “sales driven” vs. “self-service”..... Giving the right amount of credit to the effort of sales in a PLG world is much harder.

“The self serve, product, free trial motion become additional components of your attribution data.? Much like “content”, product/trial/self-serve isn’t a source, it’s a medium.? The source is still search/referral/webinar, product/trial/self-serve is the offer or content that brought them in.? Many people get this wrong.? Source and offer/content are orthogonal and mutually exclusive, they are different dimensions to attribution.”? Aaron Bird , CEO of Inflection (and founder of Bizible)

7)? Product plays a much more critical role in the funnel

• In non-PLG the entire funnel (from web visit → MQL → opp → closed won) is owned by the GTM teams.? Product isn’t involved.
• But in the PLG world, product is typically instrumental in your funnel.? They have different motivations and execution timelines. ? Sometimes you will see marketing (or RevOps) having engineering resources for certain parts of the funnel they can optimize / test more rapidly than product.