A Computed attribute is a Candidate Attribute whose value comes from a formula over your Number attributes. Talkpush calculates it for you and keeps it current, so you can combine several scores into one weighted result without spreadsheets or manual updates.
This matters because an autoflow condition compares one attribute to one value. With a single combined score, one condition can sort, shortlist, or message candidates based on several results at once. You can also filter and export by the score and use it as a message token. For more on conditions, see Autoflow Triggers.
Common Uses
Every candidate arrives with numbers on their profile: assessment results, prescreening answers, and TalkScore. A Computed attribute turns those numbers into one score you can compare and act on. Here are five ways recruiters use one.
Use case | When it helps | What you get |
Combine several results into one score | Candidates take a few different tests, and you want one number to compare them instead of reading each result. | One overall score per candidate that you can sort and filter by. |
Make some results count more | One result matters more for the role than the others, such as spoken English for a customer service job. | A score that reflects what matters most for the role. |
Catch a weak area | Every skill has to reach a minimum, and a strong average should not hide one weak skill. | A result that shows each candidate's lowest score. |
Keep "not taken" from looking like "failed" | Some assessments save a marker, such as 0 or -1, when a candidate does not finish, and a marker looks like a very low score. | Candidates who never took the test stay unscored instead of being treated as low scorers. |
Act on the score automatically | You want Talkpush to move, message, or flag candidates based on how they score. | One score that an autoflow can use in a condition, so a single rule covers several results. |
How It Works
A Computed attribute has a name, a formula, and settings for rounding and missing values. Talkpush automatically recalculates the result whenever one of the attributes the formula reads changes.
Create a Computed Attribute
You create a Computed attribute the same way as any other Candidate Attribute. For the other attribute types, see Creating a New Candidate Attribute.
Go to Leads > Candidate Attributes.
Click Add a candidate attribute.
Under What kind of attribute?, select Computed.
Enter a name in the Name field, such as Total Score.
Under Formula, click Build formula. The formula box opens.
Build your formula in the formula box. See Build a Formula below.
Click Done under the formula box.
Click Create attribute.
Note: Computed attributes are read-only on the candidate profile. You change the value by changing the attributes in the formula.
Build a Formula
The Formula box reads Number attributes. Each attribute you add appears as a chip, and you combine them with constants, operators, and functions. Type @ and part of an attribute key to insert a Number attribute. Use the operators +, -, *, and /, and group parts of the formula with parentheses.
You may also click a function under Insert to add it at your cursor. Talkpush places the cursor inside the brackets so you can type the values, separated by commas.
WEIGHTED_AVG works differently: it opens a builder where you pair each attribute with a weight. See Build a Weighted Average below.
Function | What it does | Example |
AVG | Returns the average of the values. If any value is missing, the result stays blank. | AVG(English Score, Fluency Score, Typing Accuracy) returns 80 when the scores are 80, 70, and 90. |
SUM | Adds the values together. If any value is missing, the result stays blank. | SUM(Section 1 Score, Section 2 Score, Section 3 Score) returns 24 when the scores are 8, 7, and 9. |
MIN | Returns the lowest value. | MIN(English Score, Fluency Score, Typing Accuracy) returns 70 when the scores are 80, 70, and 90. |
MAX | Returns the highest value. | MAX(English Score, Fluency Score, Typing Accuracy) returns 90 when the scores are 80, 70, and 90. |
ROUND | Rounds a value to the number of decimal places you give it, inside the formula. It works separately from the Decimal places setting, which rounds only the final result. | ROUND(Total Points / Question Count, 1) returns 7.8 when Total Points is 47 and Question Count is 6. |
ABS | Returns a value without its minus sign, so you get the size of a difference. | ABS(Interview Score - Assessment Score) returns 3 when the scores are 6 and 9. |
WEIGHTED_AVG | Returns the weighted average of the attributes you pair with weights. Weights do not need to add up to 1. If a value is missing, Talkpush leaves it out and re-weights the rest. | WEIGHTED_AVG with English Score at weight 0.6, Fluency Score at 0.3, and Typing Accuracy at 0.1 returns 78 when the scores are 80, 70, and 90. If Typing Accuracy is missing, it returns 76.67. |
Build a Weighted Average
Use WEIGHTED_AVG when some inputs should count more than others. Instead of typing weights into the formula, you pick each attribute and its weight in a builder.
Under Insert, click WEIGHTED_AVG. The Weighted average builder opens.
In the first Attribute row, select Pick an attribute and choose a Number attribute.
Enter a number in the Weight field of the same row.
Repeat for each attribute. Click Add input for another row, or click the X at the end of a row to remove it.
Click Use these inputs. Talkpush adds the weighted average to your formula.
Weights are flexible. Weights do not need to add up to 1, so 60, 30, and 10 give the same result as 0.6, 0.3, and 0.1. A weight must be a number, not an expression.
Missing inputs. When an input is missing, Talkpush leaves it out and re-weights the others.
Use or discard first. Use these inputs stays unavailable until every row has an attribute, and a message such as "Row 1 needs an attribute" tells you which row to fix. Discard closes the builder without adding anything. Done stays unavailable until you use or discard the builder.
Choose Rounding and Missing Values
Some assessments return a marker value, such as -1 or 0, when a candidate was not scored. Without a setting for it, a marker looks like a very low score. Use these three fields to control rounding and keep markers out of the result.
Setting | What it does |
Decimal places | Choose 0 to 4. The default is 2. Talkpush rounds the final result once, rounding half up, and always shows that many digits, so 70.4 shows as 70.40. |
Missing if ≤ | Treats any value at or below the number you enter as missing. Leave it empty to apply no limit. Enter 0 to ignore 0 and every negative value. |
Also treat as missing | Treats specific values as missing. Type each number and press Enter to add it as a chip. You can add up to 20 numbers, such as -1 and 999. |
A value counts as missing when it is blank, is not a plain number, is at or below Missing if ≤, or matches a number in Also treat as missing. What happens next depends on where the value is used.
Where the value is used | What happens when it is missing |
Inside WEIGHTED_AVG | Talkpush drops the value and scales the remaining weights so they add up to 1. If every value is missing, the result is blank. |
Anywhere else in the formula | One missing value makes the whole result blank. This applies to AVG, SUM, and plain arithmetic such as + and *. |
Check Your Formula Before You Save
Two checks below the Formula box show whether your formula works before you click Done.
Valid message. Valid and the number of attributes the formula reads, such as "Valid · reads 3 attributes", confirm that the formula works and that it reads the attributes you expect.
Preview with sample values. Talkpush takes a recent application that has a value for every attribute in the formula, lists those values, and shows the result on the right, rounded to your Decimal places.
No result yet. If no recent application has a value for every attribute in the formula, the preview shows a dash and says so. The formula can still be valid. Save it and check a real candidate profile after applications come in.
Check the math. Compare the preview with a calculation you do by hand before you click Done.
Choose Who Sees the Attribute and Where
The Who sees it, and where section has switches that control where the attribute appears. Two more switches sit under Owner-only & compliance, which you open with the arrow. All of them are optional, so you can create the attribute without changing any.
Setting | What it does |
Show as a column in the leads list | Makes the attribute available in Edit columns, so you can show it as a column in the leads list. It does not change who can see the attribute. |
Add to every new candidate’s profile | Adds the attribute to the profile of every new candidate. It shows up empty until it has a value. |
Sync across a candidate’s linked applications | Not available for Computed attributes. The switch stays off. |
Hide from shared guest profiles | Hides the attribute when a candidate’s profile is shared with guests. |
Read-only on the candidate profile | Always on for Computed attributes. The value comes from the formula, so you cannot edit it on the profile. |
Visible to company owners only | Under Owner-only & compliance. Only company owners see the attribute. Other managers never see it. |
Hidden everywhere (compliance) | Under Owner-only & compliance. Talkpush still collects the attribute, but removes it from profiles, exports, and filters. It also turns off the leads-list column. |
Note: If you turn on Hidden everywhere (compliance), you cannot filter or export by the attribute, and it no longer appears as a column in the leads list.
Example: A Weighted Total Score
Say you score candidates on spoken English, fluency, and typing accuracy, and you want spoken English to count the most. Create a Computed attribute with these settings.
Field | Value |
Name | Total Score |
Formula | English Score * 0.6 + Fluency Score * 0.30 + Typing Accuracy * 0.10 |
Decimal places | 2 |
Missing if ≤ | 0, so a not-scored marker never counts as a low score |
A candidate with an English Score of 80, a Fluency Score of 70, and a Typing Accuracy of 90 gets a Total Score of 78.00.
If any of the three scores is missing, the Total Score stays blank. To get a result from the scores a candidate does have, use WEIGHTED_AVG instead.
Use and Update the Result
Automatic updates. Talkpush recalculates the value whenever one of the attributes in the formula changes.
Where you can use it. Use a Computed attribute in Autoflow Triggers, in filters, in exports, and as a message token. For tokens, see The Ultimate Token Guide.
Calculate Applications That Already Exist
When you save a Computed attribute, Talkpush calculates it for your existing applications in the background. This can take a while if you have many applications. The background calculation does not trigger autoflows, so past candidates do not receive messages because of it. Only changes that happen afterward can trigger an autoflow.
Q: Can I edit a Computed attribute's value on the candidate profile? A: No. The value is read-only and comes from the formula.
Q: Why can't I find my attribute in the formula? A: The formula reads only Number attributes. If your scores are saved as a Text attribute, change its type to Number in Leads > Candidate Attributes, after you check where else the attribute is used and that every stored value is a plain number.
Q: Which attributes can I use in a formula? A: Number attributes. They can come from prescreening questions, integrated assessments, TalkScore, or custom Number attributes.
Q: Why is the value blank on a candidate? A: The result stays blank when an attribute in the formula is blank or counts as missing, unless the formula uses WEIGHTED_AVG. Check the candidate's attributes. A Computed attribute also stays blank on older applications until the background calculation reaches them.
Q: What if 0 is a real score on my assessment? A: Leave Missing if ≤ empty. Then add only the marker values your assessment uses, such as -1, to Also treat as missing.
Q: Can a Computed attribute rank candidates against each other? A: No. It calculates one candidate at a time, so it cannot return a ranking such as the top 20 percent. |
Now that your attribute is ready, see Autoflow Triggers to use it in a condition.










