PR, seeding and influencer marketing: budget and evaluation
The document plans the budget for PR, gifting (seeding) and blogger integrations, defines how to judge the result by promo codes, brand searches and contribution to margin, and sets a procedure for reputation and crisis work. It decides whom to send goods to, how much to pay and when to stop.
Why this document
PR and blogger work look like spending without return until results are tied to sales: a gifted item costs money, and visibility does not guarantee purchases. The marketing director and the PR manager read this document when they decide whom to send gifts to and whom to pay for an integration. It shows how to trace the effect through promo codes, brand searches and margin, and the procedure answers what to do in a crisis. Without it the budget is handed out by acquaintance, and money goes to reach that does not buy.
At a glance
- What is this document and why is it needed?
- The document plans the budget for PR, gifting (seeding) and blogger integrations, defines how to judge the result by promo codes, brand searches and contribution to margin, and sets a procedure for reputation and crisis work. It decides whom to send goods to, how much to pay and when to stop.
- Who owns the document and how often is it rebuilt?
- Brand: PR manager or brand manager, a season budget and a review after each wave; the marketing director decides on the budget, the owner on a crisis. Retailer: chain marketing and head of digital; own placements and partner brand placements are kept on separate lines.
- What mistakes are the most common?
- Gifts are counted at retail price or not counted at all → the budget is distorted and ROMI overstated → include cost, shipping and manager time. The effect is assessed by reach and likes → money goes to nice but non-selling channels → set a criterion on orders, contribution and queries before launch. Direct and indirect effect are added without a control → an unrealistic figure → show layers separately and add a control period.
For whom
- Brand
- The brand pays for gifts at cost, not at retail price, counts shipping and the manager time and measures return in three layers: direct promo-code orders, growth of branded queries and followers, and the long effect on reputation. The direct effect does not show the whole result, but layers 2 and 3 cannot be passed off as proven without a test. Reputation (ORM) is kept separately: mention monitoring, a reply policy and crisis scenarios.
- Retail
- A retailer promotes the store and assortment, often jointly with partner brands: cost shares are divided by contract and promo codes differ by channel. For a retailer attracting visitors to the store (ROPO: researched online, bought offline) and a new customer matter, so visits and new buyers are assessed, not only online orders. Chain reputation includes reviews on maps and marketplaces and store replies.
What it consists of
Budget and plan
- Brand: Lines: gifts (cost), shipping, fees, production, agency commission, monitoring, legal support for labeling.
- Wave plan: goal, list of channels and creators, date, expected orders and reach, success criterion set before launch.
- Brand: waves are tied to the collection calendar and drops: gifts go out 2-3 weeks before the sales start so posts land in the launch window.
- Retailer: waves are tied to store openings, seasonal campaigns and the loyalty program; exclusive brand capsules are planned with the brand and costs are shared.
- Reserve for crisis spend and difficult cases; the reserve share is recorded in advance.
Result evaluation
- Layer 1: promo codes and links (orders, buyout, revenue, contribution, ROMI on contribution).
- Layer 2: branded queries, followers, direct traffic; uplift against the base week and a comparable period.
- Brand: layer 2 rests on queries for the brand name, followers and direct traffic to the brand site; growth is counted against the base four weeks and the comparable period last year.
- Retailer: layer 2 rests on queries for the chain name, traffic to stores and the site and the share of new customers; queries for third-party brand names are not credited to the retailer but shared with the brand by agreement.
- Layer 3: quality of mentions and negative share; the metric moves slowly and is assessed quarterly.
Reputation and crisis (ORM)
- Brand: the main review topics in monitoring are quality, fit, color and size; data goes to product and production, and a repeated defect opens work with the factory.
- Retailer: the main review topics in monitoring are delivery, service, returns and store operation; a complaint about a third-party brand product is passed to the brand under the protocol, while the retailer answers the customer within its own deadlines.
- Monitoring: brand, founder name, product, reviews on platforms; check frequency and owner.
- Brand: for a serial defect the brand itself announces the batch recall and notifies retailers under the contract; the template text and the responsible contact are prepared before the season.
- Retailer: on a brand signal or by the number of complaints the retailer pulls the item from sale, stops shipments and checks stock; responsibility and deadlines are fixed in the supply contract.
- Reply policy: terms (first reply within a working day), tone, who replies publicly, when the lawyer joins.
- Brand: Crisis scenarios: defect, a dispute with a creator, an advertising error, a data leak; steps, terms and a draft text for each.
How to read the result and which decisions follow
- Read ROMI together with the bridge: the best format is “Seeding: 40 gifts” (11.6%), the worst “Micro-influencers: 12 placements” (-63.2%); the whole wave gives −RUB 459,741 at a spend of RUB 866000. Decision: first assess payback on direct contribution, and count the halo effect (queries, followers) as a separate layer without mixing them.
- Brand: a negative ROMI of placements does not mean they must be canceled: part of the effect goes into branded queries and repeat purchases within 6 months. Decision: assess payback with repeats, but keep the share of such placements in the budget limited until the halo is confirmed.
- Retailer: influencer marketing for you is your platform brand and promo codes on your site; brands often pay for it through marketing terms. Decision: in the report show own spend and brand-reimbursed spend separately, otherwise ROMI is overstated.
- Brand: gift sends (seeding) are bought at cost but valued at retail; include the cost of the gift and delivery in spend, not the retail price.
- Retailer: a promo code and link track direct orders but not in-store purchases; for a chain add an estimate through a checkout survey or unique coupons.
Control checks
- Check: the sum of contributions RUB 406,259 minus spend RUB 866,000 equals the total −RUB 459,741 (code).
- Brand: label advertising at bloggers and file the data under the advertising law; check legal details against the current edition (unverified).
- Retailer: count buyout from actual returns, not placed orders; without buyout CPO is understated.
- Brand: fix the fee, format, term and content rights by contract before publication; without rights the material cannot be reused.
- Retailer: set the assessment period (7, 30 days, 6 months) in advance; a different period for formats makes the comparison false.
Brand and retailer: further differences
- Retailer: platform-branded promo codes and cross-promotions with a brand share spend and effect; fix the rule of order attribution so as not to count one order twice.
- Retailer: PR samples come from store stock and their write-off must be reflected in the books; for a brand a gift is a marketing expense, for a retailer a write-off or marketing expense.
Where the data comes from
| Data source | Fields |
|---|---|
| Promo codes and UTM links | Orders, buyout, average basket, new customers by promo code and link. |
| Search and web analytics | Branded queries by week, direct traffic, search by brand name. |
| Creator contracts and acts | Fee, format, term, rights, advertising identifier, reach report. |
| Mention and review monitoring | Count and tone of mentions, appeals, reply time. |
How to build it step by step
- Step 1. Define the wave goal: new audience, collection launch, winning customers back, reputation; the goal sets the metric.
- Step 2. Select creators by audience fit with the main segment and the share of real followers, not by reach.
- Brand: Step 3. Compute seeding cost: gifts at cost × number + shipping + manager time; add fees.
- Step 4. Set the success criterion before launch: orders, buyout, contribution or query uplift; record it in the plan.
- Step 5. Issue unique promo codes and links and test the code at the till and in the web shop.
- Brand: Step 6. Prepare advertising labeling: contract, identifier, mark in the post; without it the placement does not start.
- Brand: Step 7. Measure the base before the start: branded queries for 4 weeks, direct traffic, orders without promo codes.
- Step 8. After the wave compute by layer: code orders, bought-out revenue, contribution, ROMI; separately query uplift, do not add the effect to the direct one without a test.
- Step 9. Set up mention monitoring and the reply policy; assign an on-duty person and terms.
- Step 10. Review the wave: whom to continue, whom to replace; update the creator selection criterion and cost per order.
Formulas
| Measure | Formulas | Note |
|---|---|---|
| Brand: Bought-out revenue | Revenue = Orders × Basket ÷ 1.22 × Buyout % | An order by promo code without buyout does not count as money. |
| Brand: Contribution and ROMI | Contribution = Revenue × Contribution margin %; ROMI = (Contribution − Spend) ÷ Spend | Contribution margin excludes ad spend but includes logistics and commissions. |
| Brand: Cost per bought-out order | CPO = Spend ÷ (Orders × Buyout %) | Compare with the allowed acquisition cost from unit economics. |
| Brand: Branded query effect | Additional orders = (Queries after − Queries before) × Search conversion % | An assumed estimate: season and advertising can also raise queries; compare with a control period. |
Formulas in notation
- Bought-out revenue
Rev = Orders × AOV / 1.22 × Buyout%
AOV is the average ticket incl. VAT, Buyout% the buyout share as a fraction.- Contribution and ROMI
ROMI = (CM − Spend) / Spend, CM = Rev × m
m is the contribution margin; ROMI on contribution, not revenue.- Cost per bought-out order
CPO = Spend / (Orders × Buyout%)
Divide by bought-out, not placed, orders.- Extra orders from the halo effect
ΔOrders = (Q_after − Q_before) × CR_search
Q is the number of branded queries per week, CR_search search conversion.
Worked example
Exercise: three formats in one wave (conditional numbers)
| Format | Spend, RUB | Code orders | Buyout, % | Bought-out revenue, RUB | Contribution, RUB | ROMI on contribution, % | Cost per bought-out order, RUB |
|---|---|---|---|---|---|---|---|
| Seeding: 40 gifts | 86,000 | 62 | 72 | 234,177 | 96,013 | 11.6 | 1,927 |
| Micro-influencers: 12 placements | 420,000 | 118 | 66 | 376,633 | 154,419 | −63.2 | 5,393 |
| Large: 3 placements | 360,000 | 96 | 70 | 380,066 | 155,827 | −56.7 | 5,357 |
| Total | 866,000 | 276 | 990,875 | 406,259 | −53.1 |
On the direct layer the wave gives ROMI of -53.1%: seeding pays back better than large placements (11.6% against -56.7%). Layer 2: branded queries grew from 1800 to 2300 per week, at a search conversion of 3.1% that is about 15.5 extra orders per week; they are not added to the direct result without a control period. All numbers are conditional.
Figures
Data as a table
| Measure | ROMI |
|---|---|
| Seeding: 40 gifts | 11.6 |
| Micro-influencers: 12 placements | −63.2 |
| Large: 3 placements | −56.7 |
Contribution-based ROMI by format. Return per rouble of spend after buyout: seeding pays off, placements do not at the conditional numbers.
- positive
- negative
- total
Data as a table
| Step | Change | Level |
|---|---|---|
| Seeding: 40 gifts | +96 | 96 |
| Micro-influencers: 12 placements | +154 | 250 |
| Large: 3 placements | +156 | 406 |
| Wave spend | −866 | −460 |
| Result | — | −460 |
Spend and contribution per wave. The contribution of each format and the wave spend: the total shows the result of the whole program.
Data as a table
| Spend, RUB thousand | Orders by code | |
|---|---|---|
| 86 | 62 | Seeding: 40 gifts |
| 420 | 118 | Micro-influencers: 12 placements |
| 360 | 96 | Large: 3 placements |
Spend and orders by format. The cost of a bought-out order grows with the integration spend: the further right and lower the point, the more expensive the order.
Specifics by category
- Apparel
- The main seeding volume: capsules and looks that can be shown whole; the blogger receives a look set, not one item. The gift size is confirmed in advance (height and size, ideally two adjacent sizes): a wrong size produces negative in the post. For core the effect is slow and comes through repeat purchases, for fashion it is fast and short (1-2 weeks after release), so cost per bought-out order is compared within the wave window, not over a year.
- Footwear
- Footwear goes out in one size, so the exact foot length and width information are needed; a wrong last turns at once into “too tight” negative. A pair costs more than a T-shirt (assumption: 2-3 times), so there are fewer recipients and selection is stricter. Placements showing wear and aging are more convincing. The ORM scenario prepares an answer to “too tight” and “too big” with a size recalculation by foot length (EU, UK, US; give Mondopoint in mm) and exchange terms within the law.
- Accessories
- The entry threshold is low: the gift is cheap at cost and needs no size, so seeding is the most mass-scale by recipient count; gift content works well (holidays, “what to give” selections). Jewelry in precious metals has higher cost and shipment insurance, and the post states the fineness mark; the words gold and silver for costume jewelry are unacceptable as a false advertising claim. Belts and headwear need a size (girth, head circumference), bags do not.
- Kids
- The blogger shows the item on a child, so written consent of legal representatives and care with the child data are needed: personal data of children is a heightened-risk area. The decision maker is the parent; the decision window is longer and tied to peaks (school, New Year). Claims of “safe” and “hypoallergenic” are backed by TR CU 007/2011 conformity documents, otherwise they breach the advertising law. The gift size is set by height (6 cm step), age is only a guide.
- Men
- The male audience more often trusts reviews and spec comparison (material, fit, wear life) than looks; a working hypothesis to test: the “review after three months of wear” format gives fewer code orders but a higher repeat share and lower returns. So ROMI is computed on contribution including repeat purchases over 6 months, not only the first order. A discount code for core is needed less often than for women: elasticity is lower.
- Women
- Fashion content has a short life: wave dates are tied tightly to the sales start so the gift is not outside the range. Because of vanity sizing the size is confirmed separately, an error gives negative and a return. More placements with looks and selections, a higher share of negative about color and fit; ORM scenarios include “color not as in the photo” and “did not fit the size” in advance. Discount sensitivity is higher, so a gift set or early drop access is better than a promo code that dilutes the full price.
Common mistakes
- Gifts are counted at retail price or not counted at all → the budget is distorted and ROMI overstated → include cost, shipping and manager time.
- The effect is assessed by reach and likes → money goes to nice but non-selling channels → set a criterion on orders, contribution and queries before launch.
- Direct and indirect effect are added without a control → an unrealistic figure → show layers separately and add a control period.
- Advertising without labeling and a contract → a fine and removed posts → include labeling in the launch step and the contract.
- A crisis is met without a policy → the reply is late and contradictory → assign owners and prepare drafts in advance.
Russian specifics
- Advertising with creators falls under the advertising law: an “advertising” mark, an identifier and data submission are required; a Roskomnadzor order sets the identifier requirements. Rule: Law 38-FZ of 13.03.2006: Federal Law of 13.03.2006 No. 38-FZ on advertising.
- Requirements for the advertising identifier and its placement are set by a Roskomnadzor order; check the current edition and the data submission procedure before launch. Rule: Order 68 of 25.03.2025: Roskomnadzor Order of 25.03.2025 No. 68 (requirements for the advertising identifier).
- Violations in advertising carry administrative liability; see the current edition of the Administrative Offenses Code for fine amounts. Rule: Administrative Offenses Code Art. 14.3, 15.12: Administrative Offenses Code, Arts. 14.3 (advertising) and 15.12 (goods marking).
- Addresses and phones of gift recipients are personal data: obtain consent and keep the minimum necessary. Rule: Law 152-FZ of 27.07.2006: Federal Law of 27.07.2006 No. 152-FZ on personal data.