Analytics and Tracking

Source of Truth

Also called system of record, authoritative dataset

The dataset a team agrees is authoritative for a given decision, so one question has only one answer.

Quick facts: Source of Truth

Category
Analytics and Tracking
Also called
system of record, authoritative dataset
Level
Beginner
Affects
Reporting consistency, decision speed, fault detection
Where to see it
GA4, Google Ads, Meta Ads Manager, your CRM or order database, Looker Studio
In this article4
  1. What a source of truth means
  2. Why a source of truth matters
  3. Common mistakes with a source of truth
  4. How to act on it

What a source of truth means

Every business ends up with several systems describing the same activity. The ad platform reports its conversions, the analytics property reports its events, the CRM reports qualified enquiries and the accounts report money received. None of them is wrong, but they will never agree. A source of truth is the agreement about which of them settles a particular question, made before the question is asked in anger.

The important word is particular. There is rarely one dataset that should decide everything. Revenue and refunds belong in your own records. Whether a campaign is worth more budget than another usually belongs to the ad platform, because only it can compare its own auctions consistently. Website behaviour belongs to analytics. Naming a single winner across all three is what causes the trouble.

Why a source of truth matters

Without one, reporting turns into negotiation. Someone brings the flattering figure, someone else brings the unflattering one, and the meeting is spent deciding whose screen counts rather than what to do next. Decisions get postponed, and the person who shouts loudest wins by default.

It also makes faults visible. When one dataset is designated authoritative and the others are compared against it on a schedule, a sudden change in the difference is a signal that something broke. If nothing is authoritative, a broken tag simply looks like another disagreement.

Common mistakes with a source of truth

The first is choosing by convenience: whichever tool is easiest to open, or whichever shows the better result. The second is declaring one and never writing down what it actually measures, so the same dashboard means different things to different people within months.

A third is expecting the designation to close the gap. Naming your CRM authoritative does not make the ad platform agree with it; it only decides which figure goes in the board pack. The gap still needs to be measured through regular reconciliation.

How to act on it

Write a short table listing your recurring questions and, beside each, the system that answers it and how that system defines the measure. Include the filters, the time zone and whether repeat actions count. This is dull work that pays for itself the first time two people disagree.

Then make the choice visible where people read the numbers. Label each figure on the dashboard with its source, keep the ad platform’s own view for optimisation decisions, and reserve your business records for anything involving money. Consistent definitions built into reporting and dashboards mean the answer stays the same whoever opens the report, and the discussion moves on to what the number should be next quarter.

Do and do not

Do

  • Pick one authoritative dataset per question, not per company
  • Write down how that system defines the measure
  • Label every dashboard figure with its source

Do not

  • Choose whichever tool shows the better result
  • Expect the designation to close the gap
  • Answer the same question from two systems

Questions people ask about this

Should analytics or the ad platform be my source of truth?

It depends on the question. For deciding which campaign deserves more budget, the ad platform is the fairer judge, because it applies one method across its own auctions. For anything involving revenue, refunds or whether an enquiry was genuine, your own records or CRM should decide. Use both, for different questions.

Can I have more than one source of truth?

Yes, and most businesses need to. The rule is one authoritative dataset per question, not one for the whole company. Problems start when the same question, such as how many leads came in last month, gets answered from two systems in two meetings without anyone saying which was used.

Does naming a source of truth make the other numbers wrong?

No. The other systems keep measuring what they measure, and they stay useful for their own purposes. Designating one dataset only settles which figure is quoted in a decision or a report. The others remain valuable as a cross-check, and a change in the difference between them is a helpful warning sign.

Related terms

Found this useful?

Share it, or ask an AI to summarise it

Back to the glossary

Knowing the term is the easy part

Applying it to your own site and budget is the work. Book a call and I will tell you what actually applies to you.