Lucid is the leading provider of visual collaboration software that helps teams see and build the future. Its visual collaboration suite includes two products Lucidchart, an intelligent diagramming application, and Lucidspark, a virtual whiteboard application.
Peter Kirk leads the marketing ops charge as the Senior Marketing Ops Manager. He is responsible for creating and delivering leads to the sales team and scaling marketing systems. Peter works closely with teams across the business, from sales to analytics to engineering and more.
As a Product-Led Growth company, Lucid has a high volume of product data stored in a data warehouse. They needed a way to utilize and analyze this data to help sales focus their outreach on the most qualified product users.
With product data living in Snowflake, separate from other sales and marketing tools, Peter and his team leveraged internal data analysts to build a scoring model and identify product-qualified users.
The Lucid marketing team worked with internal data analysts and engineers to develop and deploy a more predictive and data-backed scoring model. The team built 20 (!) different scoring model scenarios in SQL using data from Snowflake to determine the lift in opportunity creation from each. From there, they iterated on the SQL model to determine the right numerical point values, trying to make an educated guess at the weighting. Finally, they needed to pass these values into Marketo, the system of record for marketing.
As with any lead management and scoring solution, changes need to be made to adapt to evolving business needs. For Lucid, it took 100+ hours for an analyst to review and update the model, not including meeting with other stakeholders like Peter to discuss and collaborate on the changes.
Additionally, the technical limitations of Marketo meant that there was no way to account for true decay, so the team had to set up rules that they referred to as penalty points and temporary points. These rules removed points after a certain period of time to get closer to daily decay.
Connecting the disparate systems in-house is prone to errors and downtime. When data pipelines would break, the model was using inaccurate and outdated data, and sales adoption wasn’t as high as it could have been.
The Marketo set-up wasn’t only creating technical complexities, but was also affecting the day-to-day of the Lucid sales reps.
Sales was receiving PQUs (product qualified users) in Salesforce, but could not see a score or context for why the user was qualified. They were given enrichment information in Salesforce, but had to dig into other tools to see what people were doing in the product.
With these challenges, the Lucid team saw an opportunity to provide additional product usage context and strengthen the sales and marketing relationship.
The team knew they needed to evolve their PQU motion to surface higher-quality users to sales and provide the team with more context, like actions taken in the product. Peter and his team heard about MadKudu from a current MadKudu customer and felt it was the right time to evaluate their options.
They had three key criteria in their evaluation for a better solution:
Since the team had already gone down the path of building internally, they knew they needed a solution that would save them time, headache, and resources. After evaluating MadKudu, they knew it was time to make the switch.
With MadKudu, Peter and his team built a predictive scoring model without relying on the data analyst team to identify and surface product-qualified users most likely to convert.
When a change to the model or signals (context behind the score) is needed, Peter can make edits independently without analysts or engineering in the MadKudu Data Science Studio. With marketing in control, the process of making and deploying changes takes hours. Without the MadKudu platform, changes had to be prioritized by engineering and slated into a 2-week engineering sprint.
Peter and his team set up a pilot program with a group of sales reps to test out the new scoring model before rolling it out broadly. This program enabled them to collect feedback and quickly make changes.
Sales reps shared which signals were most helpful and which weren’t. For example, reps shared that they found it valuable to know what technologies a company was using, but didn’t find capital raised as valuable in adding context to their outreach. Peter was able to take that feedback and easily implement changes himself in the MadKudu Data Science Studio to make the signals more relevant.
The sales team can also see relevant product usage data and aggregations like the number of days active in different products or the number of assets created to reach out with the right message at the right time. The team has more trust in the model, and sales and marketing alignment has improved as a result.
By shifting to a data science platform built for marketers versus relying solely on internal engineering teams to build predictive models for lead prioritization, the Lucid team has a more predictive, data-backed scoring solution that supports its product-led growth goals and saves them time and resources.