How we time logging speed across photo, search and chat

Why speed is a serious metric, not a vanity one

The most common reason people abandon a tracker isn’t a lack of knowledge — it’s friction. Every extra tap is a chance to give up. So we treat logging speed as a real, weighted criterion and measure it the way you’d measure any task time: with a stopwatch and a clear definition of “done.”

What counts as a completed log

A trial ends when a correct entry is saved to the day’s diary — not when a search result appears. If an app returns the wrong food and we have to correct it, that correction time counts. This rewards apps that get it right the first time, not just apps that open a camera quickly.

The three logging methods

Where available, we time each app on three flows:

• Photo — capture a meal and confirm the entry
• Search / manual — find and add a food from the database
• Chat or voice — describe the meal in natural language

Each task is repeated across multiple trials on the same device and averaged, with outliers (a mistyped search, a network stall) re-run.

What we found

Chat and photo flows consistently beat traditional database search. The fastest app averaged about 2.6 seconds per meal via chat; database-first apps were several times slower for the same meal because search-and-confirm has more steps. The apps that were quickest to a correct log also held users best across our two-week diaries — evidence that reducing friction is one of the highest-leverage things a tracker can do.

Limitations

Times depend on device, network and the specific foods chosen, so we hold those constant within a test round rather than comparing across rounds. We report relative standings, not absolute guarantees of how fast your phone will be.