Agency Stack Recommendations
Starter Agency (1-5 clients, $500-1,000/month tooling)
Clay Starter ($134/month) for enrichment and ICP-specific workflows + Instantly Growth ($37-97/month) for email sending with unlimited warmup. Total: ~$170-230/month. This handles 5 client campaigns with moderate volumes (5,000-25,000 emails/month total).
Growth Agency (5-15 clients, $1,000-2,500/month tooling)
Clay Growth ($446/month) for advanced enrichment with more credits + Smartlead Custom ($174/month) for white-label client management + Apollo Basic ($49/month) for database access. Total: ~$670/month. Add Lavender Teams ($69/user/month) for email coaching across your team.
Scale Agency (15+ clients, $2,500+/month tooling)
Clay Enterprise with custom credits + Smartlead Custom or Instantly Hypergrowth for volume + Reply.io Agency ($139/user/month) with Jason AI for autonomous campaign management. At this scale, Reply.io's AI SDR agent (Jason) can autonomously build and manage campaigns, reducing the per-campaign labor cost.
Affiliate Revenue Opportunities for Agencies
Smart agencies generate additional revenue by recommending tools to clients through affiliate programs:
- Instantly: 20-40% lifetime recurring — a client paying $97/month generates $19-39/month in perpetuity
- Smartlead: 15-35% tiered — higher commissions as you refer more clients
- Seamless.AI: Up to 40% first-year — strongest one-time commission
- Apollo: 15-20% first-year — solid for all-in-one recommendations
- Cognism: 20-25% recurring — excellent for European-focused clients
An agency managing 20 clients and recommending tools through affiliate links can generate $500-2,000/month in passive affiliate revenue on top of service fees.
Why Agency Sales Tooling Is Different
In-house sales teams buy tools for one company, one ICP, and one brand. Agencies buy tools for ten companies, ten different ICPs, and ten different brands — often simultaneously. This single structural difference cascades into almost every other decision an agency makes about its tech stack, and it's why most "best-of" lists written for SaaS SDR teams are actively misleading when applied to agencies. What works brilliantly for a 20-person in-house team can bankrupt a 5-person agency managing 15 accounts.
The first difference is multi-client isolation. An agency can never let Client A's prospecting data leak into Client B's workflows, either legally (data processing agreements) or practically (nobody wants their competitor's leads mixed with theirs). This rules out tools built around a single shared workspace and strongly favors platforms with true sub-account or workspace isolation, like Smartlead, Instantly, and Clay. Tools like Apollo and Outreach were designed for single-tenant use and require workarounds (separate logins, separate seats per client) that break down at scale.
The second difference is white-label presentation. Agencies frequently need to deliver reports, dashboards, or even live tool access to clients under the agency's own brand — not the vendor's. This is table-stakes for retained-services agencies that want to avoid training clients to bypass them and buy direct. Smartlead leads the market on true white-labeling; Instantly and Clay offer partial capabilities. Most other tools offer no white-labeling at all, which forces agencies into "black box" delivery models (PDFs, Loom videos) instead of live dashboards.
The third difference is billing complexity. An agency paying $446/month for Clay Growth needs to allocate that cost across clients correctly — is it per-client, per-seat, or shared overhead? Tools with usage-based pricing (Clay credits, Apollo data credits) make this particularly tricky because the unit of consumption doesn't map cleanly to a single client. Smart agencies build a simple allocation spreadsheet at setup time and review it quarterly; agencies that don't end up silently subsidizing unprofitable clients with margin from profitable ones.
Multi-Tenancy Considerations
The multi-tenancy question is really about how cleanly a tool separates one client's work from another. Three models are worth understanding, because each has different implications for agency operations, data safety, and margin.
The sub-account model (Smartlead, Instantly) gives each client their own isolated workspace under a single master agency account. The agency logs into a parent dashboard, sees all clients, and can drill into any one of them. Data is strictly isolated; billing rolls up to the agency; client-specific reporting comes for free. This is the cleanest model and the one to default to whenever possible.
The shared-workspace model (Apollo, Outreach) puts all clients into one workspace with tags, folders, or custom fields to distinguish them. It's fine for 1-3 clients but falls apart around 5-10: sequence templates collide, reporting requires manual filtering, and there's a real risk of accidentally sending Client A's campaign to Client B's list. Agencies using this model usually end up running separate subscriptions per client, which destroys margin.
The workflow-composition model (Clay) is a hybrid. Clay lets you duplicate workflows per client while sharing the underlying credit pool and enrichment providers. This is ideal for agencies that want to build a library of reusable enrichment recipes and clone them per client ICP. The trade-off is that credit consumption is harder to allocate cleanly — you need a discipline of labeling workflows by client from day one.
Margin Math: What Tools to Resell vs Absorb
Agency profitability lives and dies on this question: which tool costs do you pass through to the client (resell) and which do you bury inside your service fee (absorb)? Get it wrong and you either price yourself out of deals or silently bleed margin on every account.
The simple rule is: absorb tools that deliver agency productivity; resell tools that deliver client-visible value. Clay, Apollo, and Lusha are productivity tools from the client's perspective — they help your team work faster, but the client doesn't care how the list gets built. Bury these in your fee and price for the outcome (leads, meetings, pipeline). Conversely, Smartlead and Instantly are client-visible infrastructure: the domains, mailboxes, and inbox rotation often get set up on the client's behalf and remain theirs after the engagement ends. These are legitimate to resell at cost or with a small markup, because the client gets a tangible asset.
The classic margin killer is absorbing a variable-cost tool with unpredictable usage. Clay credits are the canonical example: a client with a large ICP can burn 10x the credits of a smaller client in the same month, and if you're absorbing the cost, your margin evaporates on the big account. The fix is either tiered pricing by ICP size or an explicit credit pass-through above a threshold. Forward-thinking agencies write this into their MSAs.
As a rough 2026 benchmark, a healthy agency runs tooling costs at 10-20% of client revenue. If you're charging $3,000/month and spending more than $600/month on tools for that client, something is wrong — either your pricing is too low or your stack is too heavy.
Agency Use Cases by Tool Category
Different categories of tools serve different agency service lines. Understanding which category solves which problem helps agencies avoid over-buying (paying for features they won't use) and under-buying (missing capabilities that clients expect).
For data and enrichment, agencies building custom ICP workflows should start with Clay and supplement with Apollo or Lusha as underlying data sources. Clay's waterfall model is essentially designed for agency use: build once, clone per client, adjust the enrichment sequence for each ICP. For high-volume but simpler ICPs, Apollo alone is often enough and costs less.
For cold email infrastructure, Smartlead is the default for agencies that need true white-label and client sub-accounts. Instantly is the better choice for agencies that prioritize raw sending volume and deliverability features over white-labeling. Both tools have displaced Outreach and Salesloft in the agency market almost entirely, because Outreach/Salesloft per-seat pricing doesn't scale down to agency economics.
For AI SDR agents, Reply.io's Jason AI has become the go-to for agencies that want to run managed campaigns without scaling human headcount. A Jason-powered campaign can run mostly autonomously after setup, which means an agency can take on incremental clients without proportional hiring. The trade-off is less customization than a human-run campaign.
For coaching and quality assurance, Lavender's real-time email scoring is useful for training junior SDRs on the agency team, especially when multiple SDRs write copy for multiple clients with different voices. Gong is almost never worth it for agencies — the pricing and setup overhead don't fit the agency cost structure.
Common Agency Pitfalls
The most common pitfall is buying enterprise tools for agency workflows. Outreach, Salesloft, ZoomInfo, and Gong are all phenomenal platforms — for in-house teams at $100M+ ARR companies. For agencies, their per-seat pricing models and single-workspace architectures create cost and operational nightmares. Agencies that fall into this trap usually discover it 12 months in, at renewal, when the math suddenly stops working.
The second pitfall is failing to allocate tool costs per client. Agencies that charge flat monthly fees and lump all tooling into "overhead" end up subsidizing their least-profitable accounts with margin from their best ones. The fix is a simple per-client P&L with allocated tooling costs, reviewed quarterly. Agencies that do this consistently outperform those that don't by 10-20 points of EBITDA margin.
The third pitfall is ignoring deliverability until it's too late. Agencies running cold email at scale will eventually have a domain or mailbox blacklisted — it's a matter of when, not if. Smart agencies build deliverability into the service from day one: multiple warmed-up sending domains per client, explicit bounce-rate SLAs, and monthly domain reputation reviews. Agencies that treat deliverability as an afterthought eventually lose clients when a campaign silently stops landing in inboxes.
The fourth pitfall is over-indexing on tool recommendations over execution quality. Clients don't pay agencies for tool selection — they pay for meetings, pipeline, and revenue. An agency obsessing over which tool to add next is usually avoiding the harder conversation about why their current campaigns aren't performing. The tools matter, but they matter less than ICP clarity, offer positioning, and message quality.
Tools to Avoid for Agencies
Outreach and Salesloft require per-seat pricing with high minimums — cost-prohibitive for agencies managing multiple small clients. Gong is designed for in-house teams, not agency workflows, and its pricing structure assumes a single-company use case. 6sense and Demandbase are enterprise ABM tools with $40K+ annual commitments — they don't fit the agency model unless you're delivering enterprise ABM as a service, which few agencies do. ZoomInfo is similarly locked into single-tenant enterprise pricing and multi-year contracts that don't flex with agency client churn.