How to compare what was said vs. what actually happened
When you have a nagging sense that someone's words and actions don't line up, the worst thing you can do is rely on memory alone. Memory is reconstructive. It shifts with mood, time, and repetition. A better approach is to build a structured comparison using the records you already have - messages, emails, texts - and match them against what actually took place.
This method works whether you're dealing with a co-parent who agrees to pickup times and then doesn't show, a contractor who promised deliverables by a certain date, a colleague who committed to a task in writing, or a partner whose assurances never seem to materialize. The process is the same: capture what was said, note what happened, and let the record speak.
Why written records matter more than recollection
Most disagreements about what was promised come down to competing memories. "I never said that" is one of the most common phrases in any dispute - personal or professional. The problem is that both parties may believe their version. Memory isn't a recording. It's a reconstruction influenced by emotions, subsequent conversations, and how much time has passed.
Text messages, emails, and chat logs don't have that problem. They say what they say. A message sent on March 3rd saying "I'll have the draft to you by Friday" either exists or it doesn't. The draft either arrived by Friday or it didn't. When you have both data points, the comparison makes itself.
The two-column method
The simplest framework for this kind of analysis is a two-column log. On the left: what was said, with the date and source. On the right: what happened, with the date and any supporting evidence.
For example:
| What was said | What happened |
|---|---|
| March 3, text: "I'll pick up the kids at 5pm on Saturday" | March 7: No show. Kids were picked up at 7:45pm after three follow-up texts |
| April 12, email: "The revised invoice reflects the agreed scope" | April 12: Invoice includes two line items never discussed or approved |
| June 1, Slack message: "I'll handle the client call tomorrow" | June 2: No call was made. Client followed up asking why no one reached out |
The power of this format is its neutrality. You're not interpreting motives. You're not assigning blame. You're placing two facts side by side and letting the contrast do the work.
How to pull the records
Start with searchable platforms. Most messaging apps let you search by keyword, date, or contact. If you're looking for commitments, search for phrases like "I will," "I promise," "by Friday," "I'll handle," or "don't worry about it." These are the linguistic markers of commitments.
For each result, note the date, the exact phrasing, and the platform. Then in a separate column or document, record the corresponding outcome. Did the thing happen? When? Was there a follow-up conversation about why it didn't?
If you're working with a large message history, don't try to catalog every exchange. Focus on specific topics or time periods where you noticed a gap between words and actions. A targeted review is more useful than an exhaustive one.
Keeping the log factual
The temptation is to editorialize. "He said he'd be there at 5 but OF COURSE he wasn't" doesn't serve you if you ever need to show this record to a mediator, lawyer, or HR representative. Stick to observable facts. Dates, times, direct quotes, and outcomes.
A good test: could someone with no context read this entry and understand exactly what happened without needing your interpretation? If yes, the entry is well-written. If they'd need your emotional framing to understand the significance, revise it.
This discipline also protects you from your own biases. When you commit to recording only what is verifiable, you create a document that holds up under scrutiny - including your own.
When the pattern becomes the point
A single broken promise might be forgetfulness. Two might be coincidence. But when the two-column log fills up and the same pattern repeats - commitments made, commitments not kept, sometimes with apologies that themselves become a pattern - the record tells a story that individual incidents can't.
This is where the method becomes most useful. Not as a gotcha for any single instance, but as a way to see the shape of something over time. Patterns are harder to dismiss than isolated events. They're also harder to gaslight away, because the record exists independent of anyone's memory or characterization.
Whether you're preparing for a custody mediation, building a case for a contract dispute, or just trying to answer the question "am I imagining this?" - a factual comparison of what was said against what happened gives you something solid to work with. Not an opinion. Not a feeling. A record.