How to pick a martech stack for a two-person team that supports privacy-first retargeting under £5k/year

How to pick a martech stack for a two-person team that supports privacy-first retargeting under £5k/year

I’ve built and advised small teams on tight budgets enough times to know the pressure: you need measurable growth, modern capabilities like retargeting, and you must respect privacy and compliance — all without ballooning costs or hiring an engineering team. Here’s a pragmatic way to pick a martech stack for a two-person team that supports privacy-first retargeting for under £5k/year.

Define the problem, not the tech

Before you buy anything, get clear on what you actually need. For two people this usually collapses to three priorities:

  • Capture and store consented audience signals with minimal engineering overhead.
  • Run effective retargeting campaigns across key channels (Google, Meta, and one or two programmatic partners) without leaking PII.
  • Measure performance and attribute conversions in a privacy-compliant way.
  • If you can summarise your needs into these outcomes, you avoid feature bloat and keep costs predictable.

    Privacy-first principles to apply

    For retargeting that’s privacy-first, I lean on these rules:

  • Prefer hashed or aggregated identifiers over raw PII.
  • Collect consent explicitly and store it with timestamps.
  • Minimise data retention and segment at the edge (in the browser or client) where possible.
  • Use server-side APIs only for hashed audiences or event aggregates.
  • These rules narrow the tools you should consider: look for consent management, first-party data capture, and audience syncs that accept hashed lists or rely on privacy-preserving APIs like Google’s Enhanced Conversions or server-side conversions for Meta.

    Core components your two-person team needs

    In practice you’ll want a small set of tools that each do one job well:

  • Consent & preference management: Capture consent and record it against events.
  • First-party data capture & CDP-lite: Collect, unify, and persist visitor signals (email hashed, event stream).
  • Audience activation: Push audiences to ad platforms via privacy-preserving APIs or hashed lists.
  • Analytics & attribution: Measure conversions and build simple attribution models without relying on cross-site identifiers.
  • Creative/ad ops lite: Simple ad creation and scheduling—no complex DSP unless needed.
  • My recommended stack under £5k/year

    Below is a practical stack that I’ve tested in lean teams. I pick tools that are easy to maintain, have clear privacy features, and integrate well.

    LayerToolEstimated annual cost (GBP)Why
    Consent & CMPCookiebot or CookieYes£0–£300Simple compliance, straightforward APIs to signal consent to other tools
    First-party data capture / CDP-liteSegment (Twilio Segment) or Fathom/ Plausible with server-side storage; or PostHog Cloud (low tier)£0–£1,500Central event hub, identity resolution, can forward events server-side to ad platforms
    Server-side tagging / audience syncGoogle Tag Manager Server or a modest AWS Lambda + API Gateway endpoint£0–£300Prevents leaking client identifiers; enables hashed conversions, Enhanced Conversions
    Ad platforms / activationMeta Conversions API, Google Enhanced Conversions, and Google Ads Customer Match£0No software cost — activation via APIs but advertising spend is separate
    Analytics & attributionLooker Studio (free) + GA4 or PostHog£0–£300Free/low-cost reporting with privacy-conscious data models
    Ad creative / managementCanva Pro + Meta Ads Manager£100–£200Low-cost creative and direct campaign management

    Estimated total: ~£400–£2,600/year excluding ad spend. This leaves room for a small DSP trial or a paid Segment plan if you scale up usage.

    How to wire these up — step-by-step

    Quick setup sequence for two people, focused on minimal engineering:

  • Install a lightweight CMP (Cookiebot/CookieYes). Configure categories and make sure the CMP exposes consent state to the page (dataLayer or JS window).
  • Set up a server-side event collector. I prefer Google Tag Manager (server) hosted on a cheap Cloud Run or similar; alternatively, a small Node/AWS Lambda with an endpoint to receive browser events that have consent attached.
  • Use your CDP-lite (Segment/PostHog). Send first-party events from the server endpoint. Ensure you only send hashed emails (SHA256) or anonymised identifiers.
  • Activate conversion APIs: connect your server-side events to Meta Conversions API and Google Enhanced Conversions. Both accept hashed emails or event-level data and respect consent because your server checks the CMP state before forwarding.
  • Set up audience exports: create hashed customer lists or audience segments in your CDP and push to Google Customer Match or Meta as hashed lists. Keep retention short (30–90 days) for retargeting segments.
  • Measure with GA4 / Looker Studio or PostHog. Use event-driven tracking and link conversions to ads via the platform’s server-side features so you’re not dependent on third-party cookies.
  • Trade-offs and things I’ve learned

    In small teams every decision has trade-offs. These are the ones I expect you’ll face:

  • Simplicity vs. control: Using server-side GTM and Segment reduces client-side complexity but adds a small infra and maintenance overhead. For a two-person team this is manageable if you automate deployments.
  • Hashing vs. deterministic IDs: Hashing emails for audience uploads is safer, but you lose some match rates versus raw PII. Expect slightly lower match rates but much better privacy posture.
  • DIY vs. managed CDP: You can stitch together PostHog + server endpoint to save money, but paid CDPs give cleaner activation channels. My preference: start cheap, then upgrade when you hit scaling pain.
  • Attribution accuracy: Privacy-first approaches make last-click and cross-device attribution harder. Rely on incrementality tests (holdouts) and aggregated conversion modelling more than deterministic user paths.
  • Operational tips for a two-person team

    Operational discipline keeps this stack manageable:

  • Document every event you send (schema, required fields, consent check) in a simple events registry (Google Sheet or markdown).
  • Automate deployments and tests for the server endpoint. A single bad change can break all conversions.
  • Keep retention policies short by default and review monthly.
  • Make audience naming and lifecycle rules explicit so campaigns aren’t targeting stale segments.
  • Quick checklist before you launch

  • Consent CMP in place and returning a reliable consent state.
  • Server-side collector receiving events and verifying consent.
  • Events hashed/anonymised where required.
  • Conversions wired to Meta/Google via their server APIs.
  • Audience retention and deletion policies defined.
  • Reporting dashboards connected and sanity-checked with ad platform conversions.
  • Privacy-first retargeting on a shoestring budget is very achievable. The trick is to pick tools that let you centralise events, forward only what’s necessary in a hashed or aggregate form, and rely on platform APIs for activation rather than client-side cookie hacks. Do that and you can run compliant, effective retargeting from a lean two-person setup for well under £5k/year — leaving budget for actual ad spend where it matters.


    You should also check the following news:

    Analytics

    How to stitch server-side events to recover missing ga4 revenue for subscription products

    14/09/2026

    I ran into this problem last year while auditing a mid-market SaaS client: GA4 was underreporting revenue for subscription products by a worrying...

    Read more...
    How to stitch server-side events to recover missing ga4 revenue for subscription products
    Martech

    How to build a privacy-first first-party attribution layer using server-side events and a low-cost data warehouse

    29/08/2026

    I want to walk you through a practical, privacy-first approach to building a first-party attribution layer. Over the last few years I've helped teams...

    Read more...
    How to build a privacy-first first-party attribution layer using server-side events and a low-cost data warehouse