Contractor: how to choose and not lose

For eight years in performance, I worked with agencies (internally and on the client side), both with freelancers and in-house teams. All three formats make sense - but in different ways situations. The main mistake of most businesses is to choose the “as usual” format and not “as needed now.” This chapter is about choosing a contractor and how not to leak budget at this stage.

agency vs freelancer vs inhouse: when what?

The basic logic of choice is through the volume of tasks, budget, speed and level of control.

• Freelancer. Marketing budget up to RUB 200,000/month. One or two channel. Need speed and low overhead. Minus - single point of failure: gone on vacation or sick - the campaigns stopped.
• Agency. Budget from 400,000 ₽/month. Multiple channels at the same time. Expertise in narrow areas is required (for example, a separate specialist in creatives, separate for analytics). Minus - slower in communication, higher overheads (10–20% on top for management).
• Inhouse. Budget from 1,000,000 ₽/month is stable. Performance — a strategic channel, not a temporary experiment. Requires a maximum depth of product and audience. Disadvantage: fixed costs, slow hiring.

A detailed analysis of “agency versus freelancer” - separate blog article.

contractor selection checklist

Not “they have a beautiful website” and “they name 50 cases at the first meeting.” Minimum checklist:

1. 3 cases in your or related niche. They didn’t “do it to everyone,” but specifically to your segment. If there is none, there is a high risk of learning from your budget.
2. Contacts of two clients for recommendations. Real people who You can call and ask uncomfortable questions.
3. Transparent payment scheme. Not % of the advertising budget (this is conflict of interest - it is profitable to spend more). Fixed rate or mixed (fix + bonus for results).
4. Regular reporting. Weekly or at least bi-weekly report. Not “a month of silence and a general report.”
5. You still have access to your accounts. Not a single campaign in their account - only in yours, with guest access for them. Otherwise, when you break up You will lose all data and history.
6. Understanding unit economics. On the first call they should ask about LTV, margin, target CAC. If they don’t ask and immediately say “we know how” - red flag.

red flags in the brief and contract

• “We guarantee X leads per month.” Leads are a joint responsibility with landing page and offer. CPA guarantees at the start without funnel audit - Either it's a scam, or they'll just reduce the quality.
• Payment 100% in advance at the start. The norm is 50% advance, 50% upon results of the month. Or hourly with transparent tracking.
• A contract without a penalty for breaking on the part of the contractor. If they They can leave without consequences - they will leave at the most inopportune moment.
• “We will advertise in all channels at once.” No audit of current state, without prioritization - it’s “let’s try everything, something will work out.” Budget will disperse.
• One specialist for 30 clients. Ask how many are active projects from your manager. More than 5–7 - he physically cannot think about you business deeply.
• Lack of certificates/partner statuses. Yandex partner, Metrica certificates are not “beauty icons”, they are a basic filter.

how to structure the first month with a contractor

Not “here’s your budget, we’ll talk in a month.” We structure it like this:

1. Week 1. Briefing + audit of the current state (offices, landing pages, analytics). Joint document with plan.
2. Week 2. Launching test campaigns with a reduced budget (50% of planned). The goal is to collect the first data and understand CPL.
3. Week 3. Interim report. Analysis of the first data, adjustment hypotheses, reaching the full budget.
4. Week 4 Final report of the month. Plan-fact, conclusions, plan for next month.

In a month, you should have data to answer: “does it work or doesn’t it work?” Via three months - a clear decision: “we extend / change the contractor.”

contractor kpi: which ones are correct?

Bad KPIs: number of impressions, number of clicks, total reach. These metrics the contractor controls 100% (you can use “cheap traffic” and gain a million impressions), but they do not equal business results.

Good KPIs:

• CPL of a high-quality lead (not a “lead in the account”, but a lead that has passed the CRM filter)
• CAC at the end of the month linked to actual payments from CRM
• ROAS over a 90-day horizon (takes into account repeat purchases)
• Share of the channel in total revenue through mid-funnel attribution

Bonuses for the contractor are for the result, not for the process. For example: fix 80,000 ₽/month + 5% bonus on revenue raised through performance channels above the planned volume. This aligns interests.

how to change contractor without losses

If you decide to change, we do it carefully so as not to waste two weeks on the “transfer” affairs."

1. 30 days before the end of the contract, find a new contractor and conduct an audit.
2. 14 days in advance - ask the old one to upload all materials: creatives, texts, semantics, reports.
3. In 7 days - transfer admin access to accounts to your account (if not already).
4. On the day of transition, the new contractor gets access to the offices in “guest” mode, the old one is recalled in 1-2 days.
5. In the first month with the new one, do not turn off the dramatically successful campaigns of the old one, give new 2-3 weeks to study before changes.

checklist “how not to waste your budget on a contractor”

1. The agreement contains penalties for breaking on both sides.
2. All advertising accounts are on your accounts.
3. KPIs are tied to business results, not to intermediate metrics.
4. Weekly reporting is included in the contract.
5. Access to Ya.Metrica / GA4 / CRM is yours, not the contractor’s.
6. Transparent payment scheme, without % of the budget.
7. The first month is a test month, with the right to terminate without penalty.

in the next chapter

Chapter 15 - common errors in performance 2026. Top 10 errors that I see on audits: from wide audiences and lack of UTM to ignoring LAL and re-optimized auto strategies.