HYPOTHESIS TESTING SCRUM - CONVEYOR IN BITRIX24

case bitrix24

One of our clients requested a hypothesis testing funnel pilot. The problem was that the hypothesis work was done in excel, which made it difficult as the volume grew. Some testing tasks were done in Bitrix24 and it was convenient, but we needed a unifying entity that would make this work completely transparent. In other words, we needed a system for hypothesis testing.

THE HYPOTHESIS WAS FORMULATED AS FOLLOWS:

IF we make a hypothesis funnel in Bitrix24, categorize hypotheses, normalize time to run, test and analyze,

THEN there is an opportunity to:

  • see hypotheses on the kanban,
  • mark hypotheses with tags,
  • prioritize them correctly,
  • plan hypothesis work,
  • set tasks at each stage of hypothesis testing, according to the role model,

NOW WE WILL BE ABLE TO EFFICIENTLY TEST 30% MORE HYPOTHESES THAN WITHOUT SUCH A FUNNEL

In addition, we will be able to get a digitized pipeline of hypotheses with their KPIs, which we can display on the dashboard and see how the team works with a further transition to the normalization of tasks.

TESTING TEAM

The Marketing department team accepts the test funnel pilot. As part of the hypothesis testing process, the following roles are highlighted: 

  • Marketing manager,
  • Designer,
  • Webmaster,
  • Directologist,
  • SMM manager,
  • Analyst,
  • Sales Manager.

OPERATIONAL TASK IN A SCRUM GROUP 

For each hypothesis in the funnel, there is an operational task, which is displayed with all the information on the hypothesis.

In the scrum group it is convenient to work with weekly grow-meetings, and in the operational task it is convenient to conduct general communication to coordinate the team.

All other tasks for the roles are set automatically as subtasks of this task.

So we have all the necessary elements and tools to measure the cost of hypothesis testing work. Now we know how much it cost to draw creatives for a new test, and how much time it took to create a test campaign in the advertising cabinet. The work of the analyst to collect significant data on the result is also easy to measure.

This is priceless information for the manager who plans to do more quality results, using the resources of the team to the maximum.

The stages of the scrum team kanban are completely identical to the hypothesis testing funnel stage.

STAGES OF THE HYPOTHESIS TESTING FUNNEL

NEW HYPOTHESIS

At this stage, we gather all the ideas that arise for discussion at the team's weekly scrum. At the meeting, we determine which ideas will go to work, move them to the next stage to fix in detail the parameters of the upcoming testing, and form the tasks of the participants.

in the work

At the stage of thoughtfully recording our assumptions:
  • Hypothesis Category,
  • Hypothesis formulation in the format "We assume that if ...., then ......"
  • Assumed cost in hours to formulate the test,
  • Assumed cost in hours to formulate conclusions,
  • Test start date,
  • Test completion date,
  • Participation Roles - to set tasks for colleagues,
  • Task wording for each role,
  • Test method (HADI, SWOT, A/B test, Survey, CustDev),
  • And other parameters.

testing

The test start date determines the deadline for the team's tasks. All test preparation tasks must be posted at the transition to this stage, with a deadline on this date.

The operational task is set with a deadline on the test completion date.

Thus, within this stage, the team performs tasks by role until the experiment starts, then there is a period of data accumulation. On the test completion date, we move on to the next stage.

CONCLUSIONS

In the "conclusions" stage, the Analyst formulates conclusions regarding testing for team discussion at the weekly scrum.

Upon completion of the conclusions task, the hypothesis card enters the "Analysis of Results" stage.

analysis of results

All of the items that fall into this column are discussed by the participants at the weekly scrum. A decision is made on the further routing of the card - to implement changes or to one of the failures.

CHANGE IMPLEMENTATION

The stage is set and the task is to apply those changes that are found to be effective in the testing.

SUCCESS

At this stage, we accumulate a categorized archive of tested successful hypotheses.

FAILURE WITH CORRECTIONS

When the results of hypothesis testing are found to be unsuccessful, but there is a desire to adjust the hypothesis. At this stage it is proposed to make a copy of the hypothesis with modified wording and parameters.

FULL FAILURE

The experiment was found to be completely unsuccessful.

results

  1. After 3 weeks of using the pilot, the marketing manager told us that his team was doing 40% more than before the tool was implemented. Additionally, the manager was trained in task normalization and is now ready to test the following hypothesis:

    "If you normalize tasks for team members and prioritize tasks correctly, you can get an increase in productivity by getting the real output of specialists closer to 7.5 work hours per day."

  2. There is also a request to prepare a separate funnel for CustDev interviews.
  3. The hypothesis funnel has also attracted interest from other department heads. In particular, the head of the cold sales department has begun testing the funnel pilot and forming his requirements for functionality. The cold sales department is the company's second main "mouthpiece" and this team has more than enough hypotheses to test.
  4. A new hypothesis has been recorded that if successful elements from the CustDev interview funnel, will automatically fill the sales department's warm customer funnel, then sales in the segment will increase.

And what hypotheses do colleagues in your department have, but not written down and tested? Let's discuss!