HomeAnalytics Fundamentals › Server calls as a design constraint

Server calls as a design constraint

Purpose: Understand why you cannot track everything and how to prioritize.

Who this page is for

AudienceWhy it matters to you
Implementation engineersMust design within server call limits
AnalystsUnderstand the trade-off between data and cost

Server calls as a design constraint

Adobe Analytics charges by server calls (also called hits). This is a hard constraint that shapes your measurement strategy.

What is a server call?

One server call = one hit sent to Adobe Analytics.

Hits fire when:

Server call pricing

Pricing varies by contract, but a typical mid-market contract:

If you fire 1M calls/month, you've exceeded budget by 500K calls = $250 overage.

Why this matters

A high-traffic site can burn through server calls fast.

Example: A news site with 10M page views/month

So tracking must be deliberate. You cannot track every possible user action without exceeding budget.

Server call sources

Page view tracking (highest volume)

Event tracking

Automated tracking

Design for budget

Strategy 1: Selective tracking

Track only high-priority user actions:

Strategy 2: Throttle high-frequency events

If you track scroll depth:

Strategy 3: Server-side consolidation

Send multiple events in one hit:

// 2 hits: wasteful
sendEvent({ events: "event1" });
sendEvent({ events: "event2" });

// 1 hit: efficient
sendEvent({ events: "event1,event2" });

Strategy 4: Sample for low-value insights

If you want to track every field interaction:

Example Super's server call budget

Example Super's contract: 500K/month included

Actual usage:

They're well within budget and can add more tracking without overage.

Server call limits (hard ceiling)

Even with unlimited budget, Adobe has soft limits (they may throttle or reject excess traffic):

Avoiding overage

  1. Audit your current tracking — measure actual calls firing per day
  2. Estimate new tracking — if you add 10 new events, how many calls is that daily?
  3. Prioritize ruthlessly — track only what drives decisions
  4. Test on sample — roll out to 10% of visitors first, measure, then expand
  5. Monitor monthly — set alerts at 80% of budget

Measurement principle

Not everything that can be measured should be measured.

Choose metrics that:

Example Super should not track "timestamp of form page view" if no one uses that data. Save the calls for something that matters.

Quick navigation