Tonic.ai solves the data problem for developers. How do we test realistic test data without GDPR compliance risk?

Their pricing page features three products, each with distinct self-serve entry points, product-led growth upsell triggers, and a handoff to sales. They are a Series B startup with $47 million raised.

I am going to break down the use of their products, the Product Qualified Lead (PQL) levers, examine what works best with this pricing page and what doesn't, and review the cross-sell playbook.

Enterprise companies have massive databases with complex relationship data. A customer's record connects to an order history, which in turn connects to a shipping address. You can't simply scramble the names, or else the database will break.

This product connects to the company's real database, scans it for sensitive information, and then masks the identity. You can start testing on completely anonymized data while staying compliant.

A QA engineer or an engineering team leader would use Structural to start testing. They have a 14-day trial and connect it to their staging database. 

They use it on a few tables to prove the concept and reach the "aha" moment. After the trial, they transition to a pay-as-you-go, self-serve model or a paid tier.

The pay-as-you-go plan gets 20 tables max, at a monthly subscription, then $19/mo for each additional table. JTBD best for small teams or proof of value.

To jump into the paid Professional tier, gets unlimited tables but up to 10 TB, the hard restriction. Along with features and data sources. 

Tonic is betting that developers need to scale, have testing out on 20+ tables (or for small teams to grow), and they need to move to the Professional plan. 

The PQL signals can be as soon as the customer has been paying for extra tables on Pay-as-you-Go, and the monthly bill has been pretty high. Upsell team gets an alert to introduce what they can offer in their Professional plan with additional features, user seats, data connector, and compliance.

Start everyone on Pay-as-you-Go. When they hit high usage, upsell team contacts them. The number of database tables is a good cap for small teams, until they scale. 

Confusion between the number of tables as the initial meter lever in the Pay-as-You-Go model, then transfer over to unlimited tables, but 10 TB data cap in the Professional plan. 

The Pay-as-You-Go model is also in the FAQ section. 

This overall confusion only creates friction and indecision on how to proceed. Creates more "demos" with the sales team (who should be targeting upsell and larger customers) and results in higher CAC, only for the developers to start in the Pay-as-You-Go plan anyway.

Instead of having existing data, this product will generate a fake database.  For example, it could create 10,000 highly realistic hospital patient records with lab results and admission dates.

It's used by solo developers or AI data science teams. It gives them quick access on the Free plan, with a $10 AI credit to try. Once they get the value, and need to expand, they need to move into the Plus plan.

The PQL signal is when the Free plan's AI credit is exhausted. 

The Plus tier ($29/user/mo) offers $25 worth of AI credits. With a pay-as-you-go.

Easy to get started. Free AI credits. Try it. 

Unclear why Plus $29 per user based on the pricing page, when there are no highlighted collaboration features. 

Adding more users is also inconclusive if that increases AI credits. 

Companies can just create bunch of Free accounts with $10 AI credits to avoid paying.

Take your unstructured data containing PII and replace it with a fake equivalent so you can meet the compliance standard to push it into RAG ingestion.

The buyer is a compliance officer, ensuring that the team cleans the data before feeding it into an AI model. Textual is priced per word processed. Starting at $0.25 per 1,000 words.

The model is pay-as-you-go. With the PQL signal, when a large volume of data is processed, the upsell team will reach out to give a better rate for an annual contract.

It's unclear what is the PQL lever. Tonic's upsell team will need to determine what amount of word process threshold will be the trigger to outreach to the customer. 

Pay-as-you-go model means there is no lock-in. 

No credits to try (unclear what is the trial, despite Free Trial button) but understandable considering the customer might have an unknown volume of words to process before Textual finds the PII to block out. 

The pay-as-you-go pricing gives customers more flexibility but can also create inconsistency, which leads to risk and the potential for churn.

Companies like Tonic can't counter this with gentle reminders, such as a dashboard showing

"You have spent $95 this month, suggesting you save on the pro plan with 20% off."

Or

"You are about to hit the table limit, upgrade now to get more AI credits at a lower cost."

This is how you intervene with potential PQLs by identifying their pricing triggers and then offering a custom annual deal. Transitioning from the pay-as-you-go model to the subscription model can be effective.

Without these proactive nudges, a customer would likely continue to stay on the pay-as-you-go plan, resulting in inconsistent outputs and potential churn risk.

Tonic sells three products instead of having a unified platform. This creates confusion in bringing the experience to the forefront.

The marketing and sales team will have to rely on a lot of cross-selling and education via emails and sales calls. Because of these siloed products, each with its own trial experience, this restrict any potential for in-product discovery (or expectation set up front properly). Examples like "Do you need to redact PII from any unstructured logs attached to this data? Try Textual for free."

Tonic is selling bundles, which require speaking with a sales representative, and create a higher CAC.

With Fabricate, there's a $10 credit limit. With Structural, there is a 14-day trial. There's a limit to push the customer to upgrade.

Within these trial experiences, it's important to ensure the customer reaches the "aha" moment and uses their time efficiently. Tonic needs to create a great self-serve experience that supports this.

They also use a pay-as-you-go model after the trial ends or the free credits are used. Customers can choose to continue and pay a little more.

As customers scale, the upsell team intervenes to offer a better deal than pay-as-you-go. The pay-as-you-go option must be attractive enough to encourage use, but not so attractive that a custom annual contract or special deal isn't appealing by comparison.

Find the number of actions the customer needs to take to truly see value. It could be $10 in free credits or $50,000 worth of credits.

Once the customer sees the value, they're primed to upgrade to the Pay-as-You-Go plan. 

Continue testing your pricing with the Pay As You Go model to make a custom quote more attractive than Pay As You Go. Work on understanding the customer's threshold. 

Include in-app paywalls to do introductions and discover your other multiple product lines. 

The Product-Led Growth methodology here really requires testing the pricing lever. It's important to continue understanding where the thresholds are, when customers are most likely to upgrade, and what the killer features are that become the PQL levers for customers to want to jump to the next tier.

The Pay As You Go model gives customers flexibility, but it also introduces churn risk. There needs to be a strong incentive for customers to move to the next tier. You really have to graduate customers to a subscription tier by using dashboards, nudges, and cross-selling to find and push for PQL signals.

Currently, that's not very clear, and I'm sure the company is still trying to gain a better understanding of what this should look like before the pricing evolves.

Read more about pricing page breakdown

I write weekly about my learning on launching and leading PLG. Feel free to subscribe.

I am Gary Yau Chan. 3x Head of Growth. 2x Founder. Product Led Growth specialist. 26x hackathon winner. I write about #PLG and #BuildInPublic. Please follow me on LinkedIn, or read about what you can hire me for on my Notion page.