BehavioralStrategy.com teaching pack v1.0
Author: Jason Hreha
Source file: education/instagram-worked-decision.md
License for original text: CC BY 4.0. External sources retain their rights.
Editable text copy; Markdown headings and tables are retained.

# Instagram: A Worked Decision

> **Worked recommendation:** Test photo sharing for a defined group before committing to a new social network. Treat the missing evidence as work to do.

Complete the [student exercise](https://behavioralstrategy.com/education/instagram-decision/) before reading further. The responses below are original instructor analyses. They are not recommendations recovered from Instagram's team, and choosing the historical winner is not the assessment criterion.

## The historical outcome

The team focused on photography. Its 2011 account identifies three product problems: image appearance, upload delay and distribution to other services. A [September 2010 preview](https://techcrunch.com/2010/09/20/instagram/) confirms the prelaunch photo focus. [Evidence BS-0080; see source-notes.txt]

Krieger's [2015 retrospective](https://www.wired.com/2015/10/five-years-of-building-instagram/) also describes substantial work to support the service after launch. The history combines behavior choice, product design and execution. It supplies no controlled estimate of the contribution of any one choice.

## Response A: commit to a narrow prototype

**Decision record.** Prioritize making a chosen phone photo available to an intended audience. Initially serve people who already take photos they sometimes want to share. Build only enough to observe the whole action in a normal sharing situation. This recommendation authorizes a focused learning step; a wider launch still requires a separate decision.

**Reasoning.** The reported interest makes this a reasonable candidate to investigate. It does not show how many people have the need, how often it arises, or whether the proposed route improves their existing practice. The largest avoidable commitment would be building a network before learning whether people want another audience there.

### Compare the alternatives

This is an instructor comparison. It assigns no historical scores.

| Option | Contribution to the proposed objective | Reason to retain or defer it |
|---|---|---|
| Share a place | Gives an audience location-based information. | Retain if the intended group's real need concerns places; a general photo objective does not establish that. |
| Share a photo | Gives an audience visual information selected by the sender. | Investigate first, while testing the demand outside the original network. |
| Browse and respond | Helps an audience find value in contributions. | Necessary to some product models, but it cannot supply the contributions by itself. |
| Share through an existing destination | May reduce preparation and transfer work without requiring another network. | Keep as a serious alternative and a comparison for the prototype. |

### Define the action and its dependencies

**Proposed target:** when an intended participant independently wants to share a phone photo, they make a selected image available to their intended audience and can confirm that it arrived. The episode ends with that confirmation or with abandonment. This is an instructional definition, not an Instagram metric.

| Part of the chain | Responsible party | What to observe |
|---|---|---|
| A relevant sharing occasion arises | Sender | A real intention to share; a research instruction alone does not establish demand. |
| Select or capture the image | Sender | Whether the image and required skills are available. |
| Decide what and whom to share with | Sender | Audience choice, hesitation and deliberate refusal. |
| Prepare and transfer the image | Sender and product | Steps, errors, waiting and any supplied help. |
| Make the result accessible | Product and destination | Delivery or access confirmation, rather than a button press alone. |
| Encounter another relevant occasion | Sender and their environment | Another opportunity, a different route chosen, or an unknown opportunity state. |

This chain keeps the distinction between [capability and context](https://behavioralstrategy.com/glossary/behavior-fit-assessment/) visible. A person can know how to post and still lack a suitable audience or a usable connection. Some people may appropriately choose not to share.

### A concrete next test

The following is a **proposed teaching example**. Its participant count and duration are design choices for a small observation exercise, not historical numbers or universal validation thresholds.

| Protocol field | Proposed plan |
|---|---|
| Purpose | Find where intended users choose, complete or abandon this route during normal sharing occasions. |
| Recruitment | Invite 12 people who already sometimes share phone photos. Seek people beyond the team's friends. Record invitations, declines and eligibility; do not silently replace people who fail. |
| Exposure and context | Offer a working prototype for seven days. Let participants keep their usual tools. Observe naturally occurring use at home and away from home where those contexts are relevant. |
| Completion | Confirm that a participant-approved image is available to the chosen audience. Record attempted shares that stop earlier. |
| Support | Record every reminder, demonstration and troubleshooting step. Separate normal product support from extra research help. |
| First-use record | Keep invited, eligible, exposed, attempted and completed counts separate, with the same stated window. |
| Repeat record | For conditional repeat completion, count first completers who complete again over first completers with a documented later sharing opportunity. Report unknown opportunities and unknown completions separately, with missing-outcome bounds. If also counting later-opportunity completion across all participants, label that as a separate measure, including earlier failures. |
| Alternative | Record occasions on which the participant chooses their usual route or decides not to share. |

Use participant logs to identify possible sharing occasions, then distinguish directly observed events from later reports. A reported opportunity does not become an observed event merely because it appears in a log. Keep that distinction in the repeat-use record.

**Rules written before observation:** repair a recurring technical obstacle before interpreting abandonment as lack of demand. Reconsider the audience or route if people consistently choose a functioning alternative that better serves the same occasion. Gather more evidence when exposure, opportunity or completion cannot be determined. A promising small observation exercise can justify a larger test; it cannot establish a population-wide completion rate.

The decision after this exercise is whether to invest in a more rigorous field test and what it should compare. Use the [BMF protocol](https://behavioralstrategy.com/methodology/validate-behavior-market-fit/) to specify the operational requirements, uncertainty and pass conditions for that larger commitment. No observations or results are invented here.

## Response B: investigate the existing route first

An equally defensible response can defer the prototype. It would observe the same intended audience using the tools they already have, map the recurring difficulties, and test whether a new service is needed at all. Its strongest reason is the absence of representative usage evidence. Its cost is delaying direct experience with a possible improvement.

This response becomes weak if “more research” has no endpoint. A credible record names the observation that would justify building: for example, a repeated, consequential step that the team can remove while preserving the person's intended audience. It also names a stopping result, such as finding no meaningful improvement the proposed service can offer that population.

Response A accepts more prototype cost to obtain behavioral evidence. Response B accepts more delay to understand existing practice. Compare those commitments and their reversal conditions rather than treating agreement with history as proof of reasoning quality.

## Check your response

| A strong response does this | Repair this error |
|---|---|
| Defines an observable action for a specific person and occasion. | “Use Instagram” leaves the behavior and success point unclear. |
| Keeps posting, audience activity and product features distinct. | “Build filters” is a solution choice without a complete target action. |
| Names missing evidence and a finding that changes the decision. | Later growth cannot supply an earlier comparison that was never observed. |
| Includes the existing route and necessary support. | A test that removes competing tools or supplies unrecorded help can mislead. |
| Makes a limited next commitment. | A favorable reaction to a prototype does not settle sustained use or business viability. |

**Transfer question, invented scenario:** the next product serves employees who may share work images only with a specified internal audience. Which parts of your proposed chain remain useful, and which parts of a public-network strategy would you reconsider? Explain the change using the new condition, not a claim about Instagram's users.

## Source chronology and limits

| Record | Publication date | How it is used |
|---|---|---|
| [Product preview](https://techcrunch.com/2010/09/20/instagram/) | 20 September 2010 | Contemporaneous observation before public release. |
| [Founders' Stanford talk](https://stvp.stanford.edu/av/stanford-startup-entire-talk) and [transcript](https://stvp.stanford.edu/node/12536/printable/print) | 11 May 2011 | First-person recollection of the earlier decision. |
| [Krieger's retrospective](https://www.wired.com/2015/10/five-years-of-building-instagram/) | 7 October 2015 | Later account of operating the product. |

Sources checked **5 September 2026**. The packet's objectives, candidate comparison, protocol and assessment are instructor work. No recovered source shows the use of DRIVE, BMF or BFA, and the public history does not validate the integrated method.

*Original teaching material by Jason Hreha, version 1.0, licensed CC BY 4.0. External sources retain their own terms. See [teaching notes and reuse guidance](https://behavioralstrategy.com/education/teaching-notes/).*
