What is Clay? What can it be useful for?

Clay is a Customer Intelligence Platform that helps companies identify, understand, and connect with their ideal customers and prospects – Clay is designed as a comprehensive Customer Intelligence Platform that empowers businesses to identify and build deeper relationships with their target audience.

Key benefits of Clay include:

• Data Enrichment – automatically enriching contact and company data from various sources.
• Prospecting Automation – building targeted lists of potential customers based on specific criteria.
• Contact Database – maintaining an up-to-date database of prospects and leads
• Custom Workflows – creating automated sequences for outreach and follow-up.
• CRM Integration – syncing with existing CRM systems like Salesforce, HubSpot, etc.

Building Tables

How is a Clay table structured?

A typical Clay table includes:

• Column – these represent different data fields or attributes. Common examples include:
  • Contact information (name, email, phone)
  • Company details (company name, size, industry)
  • Custom fields for your specific use case
  • Enrichment fields that pull data from external sources
• Rows – Each row represents a record that can be an individual contact, company, or event
• Cell Types – Clay supports various cell types, including:
  • Text fields
  • Numeric fields
  • Formula fields (similar to Excel formulas)
  • Dropdown menus
  • Date fields
  • Rich text fields
• Enrichment Cells – special cells that can automatically fetch data from external sources like LinkedIn, Clearbit, or other data providers
• Formula Cells – cells that can perform calculations or manipulations on other data in your table. You can even use AI to easily create your own formulas.

The key difference between Clay tables and traditional spreadsheets is that Clay tables are “live” – they can continuously update with fresh data from external sources, run automated sequences, and trigger actions based on specific conditions.

Each column of a Clay table corresponds to a step in the sales process workflow – tables use spreadsheet rows and columns to structure information. Columns are populated through enrichment processes in which Clay catches information from external providers or processing workflows that analyze known information.

Tables can be connected to other systems – you can share data between Clay tables, to external systems like CRMs, and sequencing tools for email or LinkedIn.

How do you build a well-structured table and prospecting workflow in Clay? 

Step	Example
1: Reverse engineer your workflow design – define what you want to achieve with your workflow before you start building it. The outputs you envision should correspond to some of the columns in your table.	You want to send emails to qualified leads. Outputs you need include: Email addressNameCompany namePersonalized copy you will send
2: Identify the inputs and triggers that will enable your ideal output – consider both the data you must bring in and the different types of triggers for collecting that data. As your list grows, the connections between different types of data points should become clearer.	To identify new leads, triggers can come from a variety of places:CRM (e.g., they download a paper)A keyword monitoring system (e.g., every time a relevant new job post appears)
3: Design a manual workflow to create your ideal output – map out the discrete steps involved in identifying or creating connections between inputs, triggers, and outputs. Clay implementation will follow the same logic.	Consider the questions you would need to answer in order to think through this process logically. For example, if you have the ability to find a new job post:How would you qualify if the job post is relevant?How would you define that the company is relevant to you?How would you find employees at that company?How would you find their emails or any other data?How would you craft and personalize copy for those leads?

Prioritize effectiveness when designing a new table – expensive tools (e.g., GPT 4) yield better results, which can help you fine-tune your design and achieve the level of accuracy you need.

Once you’ve optimized your workflow, improve cost efficiency – Clay offers many levers for reducing the cost of running workflows. Experiment with using different types of models for different steps in your process (e.g., using lighter models or different tools within Clay for tasks that are easier to execute accurately) to optimize for both cost and effectiveness.

Look for ways to improve your workflow on a monthly basis – new AI tools are emerging rapidly, and solutions are becoming increasingly powerful and affordable. Regularly review your workflows to determine whether:

• They are maintaining their relevance
• They can be improved

What kinds of columns should be in a Clay table? 

The columns in your table should be specific to your workflow – some tables need purely contextual information in their columns, while others require more complex prompting or a mix of both.

Clay Table Purpose	Column Example
Identifying	For a list of leads, you might have columns for email, location, full name, Lead source, Lead score/qualification status, segment, etc.
Analyze data	To analyze a list of leads, you might have columns that indicate: – Potential value – Buying signals (hiring, expansion, etc.) – Competitive products used – Budget indicators – Number of employees – Technologies they use – Office locations – Number of employees per department – Whether they recently went through a funding round

Leverage Clay’s enrichment capabilities to gather comprehensive professional information – Clay can scrape LinkedIn profiles to retrieve “About” information, published posts, and data that can be used to provide valuable context for understanding your leads’ opinions and fit.

Every business’s workflows should be unique – while different companies might use the same input data, competitors don’t necessarily use the same tables to execute the same activities. For example, two competitors might use different versions of a complex prompt (e.g., a prompt that qualifies companies using a multi-factor scoring system that depends on factors like revenue, size, location, and technologies that they’ve used previously). Although the companies sell competing solutions, their qualification processes should reflect their unique GTM approaches and understanding of their customers.

What are the different sources of information you should leverage to build a table in Clay?

If your goal is to find leads and enrich databases, most use Clay’s native LinkedIn integration – use filters such as industry, size, and location to qualify companies before identifying individual leads. Enriching company data first increases the likelihood of identifying qualified companies and focuses investment in individual data enrichment on higher quality leads.

Clay offers powerful enrichment capabilities with their AI agent – Claygent can research any company or topic using publicly available online information. For example, if you have a contact’s email address, Claygent can automatically retrieve valuable and potentially extensive data about the individual and their company.

Common additional enrichment sources include:

• CRM data
• Apollo
• ZoomInfo
• Build With

How wide should your parameters be, and how should you scope the initial size of your list? 

Understand your target market well enough to create a tiered ICP – segment your ICP based on LTV to guide how you spend money across those groups.

List size depends on the size of your total addressable market (TAM) – if you have a massive audience, qualification is simpler and is focused on identifying the most valuable segmentation of that market. If you have a small niche audience, qualification must be more precise.

TAM Size	Parameter Guidance	Target Qualification Accuracy Rate
Large TAM	Create at least 2 ICP tiers – for example:1 tier for customers with an LTV of X1 tier for customers with an LTV of 10X  Scale investment to value – it might be worth investing in high-touch approaches (e.g., personal notes, phone calls, gift cards) for the 10X tier and more affordable methods (e.g., cold emails) for the X tier.	80%
Small TAM (or very high-value prospects)	Use multiple qualification processes – leverage different tools, data sources, and workflows to carefully score leads and companies.	99%

Keep in mind that most plans limit you to 50,000 rows per table – unless you have an enterprise plan, you might need multiple tables or an upgrade to reach out to a larger audience. If you plan to manage high volumes of outreach, upgrade your plan or ask the Clay support team if they can add extra rows to a specific table.

Test the first 10-50 rows of new workflows before scaling – to confirm that you can achieve the accuracy rate you need to hit. Since most workflows generate false positives or false negatives, it’s crucial to understand whether you need to hit an accuracy rate that’s closer to 80% or 99%. Test each time you adjust your workflow until you reach your goal.  

How do you maintain list hygiene, especially when dealing with multiple data sources? 

Program AI scrapers to fill in gaps in your data – for example, if you are trying to identify a marketing agency that offers specific services, you might read through their LinkedIn profile and then check their website before you find a page that lists their services. However, other companies might require a more complicated process. Program the scraper to follow the same steps you would manually take, and don’t skip the manual step when designing this process.

Compare the same data across multiple providers – use Clay to check if different data sources like Apollo, ZoomInfo, and Clarity return consistent information about company revenue or other key metrics. If a data provider has unreliable information that’s critical to your workflow, you might need to find a workaround.

How can you effectively filter and segment data within Clay tables?

Organize workflows to have 1 table per unique variable – for trigger-based workflows, use separate tables for triggers, companies, and leads. This structure allows you to waterfall data, with enriched data from the first tables falling through to the later tables and eliminating redundant enrichment costs.   

Consider campaign-specific tables for different outreach initiatives – for some campaigns, it’s easier to have leads/companies and content creation in separate tables.

Example: Tables for a workflow that creates fully automated AI-powered copywriting for a lead gen agency
Table	Table Purpose
Table 1 – Campaign data	Host all data about the campaign itself, including:ICP; Description of ICP; Buyer persona; Value proposition; Relevant pain points; Name.
Table 2 – Lead copy generation	Use the leads table to select the template that is most suitable for each lead, then leverage campaign data to craft ultra-personalized copy.

What automation capabilities exist for maintaining and updating tables in Clay?

Trigger-populated tables can constantly refresh and even trigger responses without waiting for manual intervention – connect a marketplace scraper like Apify or PhantomBuster—or a CRM or marketing automation platform—to create an evergreen source of new leads.

Clay’s webhooks let you receive real-time data updates – any tool or data provider that can send data via an API call can share that data with Clay. Use a webhook to receive the most updated information if data in a fielding is frequently changing. 

Enrichment and Prioritization of Contacts

How do you properly set up and run batch enrichment processes on contact data in Clay?

Leverage Clay’s waterfall enrichment feature – Clay acts as a marketplace for data providers that allows you to use credits to test different sources instead of committing to (possibly expensive) monthly subscriptions before trying them.

The multi-step waterfall enrichment process lets you increase data coverage while reducing costs – despite their claims, not all providers can give you all the data you need. Use providers in sequence to maximize the data you get from the cheapest sources. Start with the most cost-efficient provider, then use the second-cheapest provider to fill in any gaps. Continue this process until you have the information you need.

If needed, use validator tools before finalizing qualified lead lists – Clay has an email validator tool and a phone number validator tool that you can use to eliminate leads before sending them to a campaign. 

What strategies should you use to build effective scoring models for contact prioritization in Clay?

Use calculated formulas for straightforward qualification processes – Clay has a free column that lets you check if keywords appear in a text, compare numerical values, or match data across columns. These basic formulas are useful for simple qualification based on metrics like revenue, employee count, or technology usage. They also don’t require coding knowledge.

Use AI models for more complex qualification processes – if a scenario requires a nuanced understanding of language, prompt an AI model to determine this for you. For example, if you are scraping job posts, AI can read the post to understand whether a company is hiring because they’re investing in a department or replacing an employee.

Experiment with scoring models from different AI tools – try a variety of options, including well-known models like ChatGPT and alternatives like Claude 3.7, DeepSeek, and Perplexity.

How do you implement advanced segmentation based on enriched data in Clay?

Align your segmentation strategy with your business strategy – effective segmentation—whether it is manual or automated—supports key business objectives. Identify the most useful way for your GTM team to segment leads into groups, then automate that process.

How Clay can support different segmentation levers
Lever	Example
Segmentation by channel – segment leads by platform activity level to improve channel-specific outreach performance.	Reach prospects where they’re most active. For example, if a lead has fewer than 500 connections on LinkedIn, they are probably not active on that platform and cold outreach through a channel like email or a phone call would perform better.
Segmentation by company profile – utilize publicly available data to hone in on your target audience.	Claygent can scrape company websites, LinkedIn profiles, and other publicly accessible platforms for data on a scale that would otherwise be prohibitively time-consuming or expensive. For example, if your ideal customer is a marketing agency that provides lead gen services, Claygent can “read” prospect websites to determine which companies offer this service and segment them based on the outreach channel(s) they specialize in.
Segmentation by value – refine your definition of a tier 1 customer as much as possible before doubling down on targeting your most valuable customers.	Optimize resource allocation by: Creating highly qualified tier 1 / high-value audiences; Tailoring your segment investment to its potential value

Prospecting/Outreach

How do you integrate Clay with outreach platforms and CRMs? What platforms does it work best with? 

Clay has native integrations with HubSpot and Salesforce – simply add a column in your Clay workflow to receive new data. You can also create a schedule to regularly enrich the data in a Clay table using data from your CRM.

Connect with other CRMs through API calls or webhooks – whether it is through Clay’s webhooks, the CRM’s webhooks, or an API tool, it is possible to smoothly share data between Clay and most CRMs.  

Use external connections to contact leads and alert your team to important updates – Clay workflows can connect to LinkedIn sequencers or email sequencers to share data with SDRs and AEs, generate PDF reports to provide SDRs with lead information and scripting for outreach calls, and even send messages through Slack to let the team know if companies they’re targeting have hired for a relevant new position.

What features does Clay offer for personalizing outreach at scale?

Clay can personalize copy based on lead data – use AI-generated copy to send customized messaging to prospects through email or LinkedIn cold outreach.

Some sequencers can even send handwritten notes – for some companies, extremely high-value segments might justify expensive, highly personalized approaches like using sequencers to send handwritten letters.

How do you measure and optimize the performance of outreach campaigns using Clay?

Benchmark performance against industry standards – metrics like cold outreach reply rate have well-defined benchmarks that vary by industry. However, consider the downstream quality of those replies in addition to reply rate itself.

Use Clay to benchmark test different GTM approaches – Clay’s ability to manage multiple campaigns and numerous variables makes it easier to test different campaigns, value propositions, or even your ability to sell different products.  

Clay’s greatest value comes from supercharging your current tech stack – you can use Clay to unlock massive scale if you’ve already optimized your GTM strategy. For example:

• Measure time recovered/ time saved – if you can maintain the vetting and segmentation quality of manual lead identification, Clay should be able to process 100,000s of leads per day. Theoretically, you could conquer a market in days, though this isn’t realistic.
• Monitor reply sentiment – Clay can categorize whether replies to cold outreach are positive, negative, or referrals, then route them appropriately (e.g., to your Sales team).
• Measure platform-specific metrics – platforms like LinkedIn have connection rates in addition to reply rates. However, benchmarks for these metrics vary significantly by industry and target audience. Clay can help track performance that is more complicated to evaluate.

Overall 

How can you optimize your spend or reduce credit costs on Clay?

Turn tables off when creating/modifying a new table – otherwise, Clay will automatically run all the enrichment processes you’ve programmed when you create new cells. For example, if you upload a large CSV of leads, you could accidentally consume thousands of credits, resulting in significant costs.  

After confirming that Clay can effectively execute a workflow, separate tasks into discrete steps that each use their own model – a common way to waste money on Clay is to use the same model to do everything. Instead, use simpler models for “easier” tasks, such as scraping websites, and use more sophisticated models for complex analysis.

Implement early qualification filters to avoid spending on unqualified leads – in some cases, it makes sense to qualify leads using available data before putting that data through an expensive enrichment process. For example, if a lead doesn’t have an email address, and you’re running an email-only campaign, you shouldn’t invest in enriching that lead’s data.

Regularly review and update your workflows – more sophisticated and cost-effective tools continue to reduce the cost of running tables through Clay. A solution that is optimized for cost today might be considered unoptimized in a few months. 

What are the most important things to get right?

Always start with powerful models that will help you establish baseline effectiveness – confirm that you can get the result you want (and quantify its accuracy) before trying to decrease cost.

Follow industry leaders to stay up to date – highly knowledgeable industry leaders post about their personal and professional experiments with AI tools on LinkedIn and blogs. The tools and integrations Clay offers are evolving so quickly that it is crucial to remain educated on the emergence of new strategies that can help your business. 

What are common pitfalls?

Building workflows without clear objectives can become a time sink – it’s easy to fall into the rabbit hole of exploring different Clay capabilities and creative ways to get from A to B, especially if point B is not well-defined. When this happens, you’re more likely to get overwhelmed because you don’t know how to filter through the options available to you.